Home

VBT 1.10 User's Guide - Ecologia e Gestão Ambiental

1. Ctrl Enter Osciloscope Echogram Color Bar Colors Sel Map Spatial Rangen Bottom Sampling Windows Bottom Tracking Parameters Select Bottom Typing Method Configure Bottom Typing Algorithms Mark Bottom Gutput Report Filters Output Report Export Fields Output Report Format Options a Go to 1 ping x Transducer Frequency Hz OK Cancel 4530 35 e en 4 C Mr 49408 Copyright 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide Configuration gt Oscilloscope Bl Oscilloscope Options BioSonics Inc TVh C Mone 20LogH C 40 Log H Beam Type f Narrow C Wide Channel 4 lea Cc Fis Ca Ch Te Cy CE Depth x Scale Max Initial Offset Samples Per Meter Amplitude 1 Scale Mar Log Lin Min Background Color Fixed Solid Color Echogram Colors BioSonics Chapterl Getting Started You can change the display for mat of the oscilloscope window by selecting Oscilloscope in the CONFIGURATION MENU when the Os cilloscope window is active or by right clicking in the oscilloscope win dow The default settings are shown in the following figure TVG Time Varied Gain The decibel scale can be set to el ther None 20logR default or 40log R Chart Type
2. VBT User s Guide 2 Ground truth Library 2 1 2 Manually generating a ground truth library Step 1 Step 2 Step 3 Step 4 Step 5 Open the sample library data file mud120 dt4 located in the CTLG folder Activate method B4 by clicking on B4 in the toolbar or selecting Configure Bottom Typing Algorithms B4 from the ConFIGU RATION MENU Select the Info button in method window The Info windows contain information recorded by the user about this set of verified data files You may edit the text directly y Note When creating a ground truth library record 1 the order in which the feature spaces were created and 2 the data acquisition parameters in the Info section for future reference 120 kHz 0 4 ms SN cM Barwin 16 48 PL 16 MOD120DT4 red Bot SANDT20 DT4 green nbi ROCKT20 D T4 blue 13 In the method window select the New button to generate a new database The new database will be your ground truth library OT TO Gen kHz 0 4 ms SH CT Botwin 15 48 PL 16 MOD 120 DT4 red Bot SAND120 DT4 green iral ROCK120 DT4 blue 13 Activate the oscilloscope window by clicking on it and select Bottom Sampling Windows from the CONFIGURATION MENU and make sure the fields have the following values The data acquisition pulse is included in the window for reference and cannot be altered BS amp E 2005 03 0022 2 2 30 43 Copyright C 1998 2005 BioSonics In
3. Visual Bottom Typer Install Program BioSonics c Advanced Digital Hydroacoustics 2 lt Visual Soror 1 0537 VZT Welcome to the Visual Bottom Typer Install program This program allows you to install Visual Bottom Typer on your hard drive It is recommended that before proceeding you ensure that no other Windows programs are running Click Next to continue installing Visual Bottom Typer or Exit to safely cancel the installation You may resume A again at any time distant 9 B lici 3 8 Li A Py visual Bottom Typer I CARARE 3 Select I Accept You must accept the license agreement in order to install BioSonics Inc VBT SeaBed Classifier A copy of the license agreement is provided on Page x License Agreement xj isual Bottom Typer VBT TM END USER LICENSE AGREEMENT Introduction This is a legal agreement between you and BioSonics Inc A The terms of this Agreement govern your use of this software Please read this agreement carefully before vou install this package By installing this package you are agreeing to be bound by this Agreement If you do not agree to the terms of this Agreement please return to your retailer the entire package with the original wrapping for a full refund 4 Select Next to install VBT in the default directory C BioSonics VBT Select Yes if prompted to create a new directory
4. y um e 4 ro j TEN ou Lo Figure 4 4 Formation of the first bottom echo El and El Figure 4 2 2 illustrates the formation of the second bottom echo The second bottom echo is produced by a double specular reflection from a seabed and single reflection from the surface of the water For a flat bottom the specular reflection is directly related to the hardness of the bottom If the bottom is rough the second bottom echo will be weak k E per ae 2 ES A A F m P a ut a og j J A ubique quasi 1 an i E a sped 7 Al ry er a ut in m Figure 4 5 Formation of the second bottom echo E2 Method B2 calculates the energy of E1 and E2 for each ping in the unclassified data set and plots the pings accordingly where E2 is the vertical axis and El is the horizontal axis The bottom type of each ping is determined by the placement of the feature space boxes previously generated from the ground truth library 4 2 3 Method B3 First Echo Division Method B3 classifies the bottom type based on the ratio of the first and second parts of the first bottom echo E1 E1 The first part of the Ist bottom echo building up from beginning of a bottom echo to maximum 1st contains the bottom hardness signature The second part of the bottom BS amp E 2005 03 0022 2 2 30 77 Copyright 1998 2005 BioSonics Inc All Rights Reserved PRE
5. 5 9 10 Cluster Analysis PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started sections For first time users we recommend reading through the introduc tory tutorial in Section 1 5 If you are familiar with VBT we recommend practicing with the tutorials in Chapter 2 Advanced users are referred to Chapter 3 for information on maximizing the accuracy of your bottom typing results 1 2 Data Collection Guidelines In order to determine the bottom type using VBT the user must first collect a standard library of hydroacoustical signals while simultaneously physically determining the bottom type The physical identification estab lishes a ground truth library which will be used as the gold standard to classify the bottom type encoded in subsequent echo signals It must always be kept in mind that software based classification is the final step in a field methodology that takes careful planning and execution During data acquisition please keep in mind the following guidelines e Data collection with a factory calibrated echosounder e Deploy the transducer from a stable towed body Use an orientation sensor to measure off axis pitch roll and yawn and correct for any deviations e Maintain a boat speed that minimizes off axis deviations e Collect data during calm weather e Conduct a preliminary bathymetric study of the survey area and record bathymetric gradient locations e Lay transects
6. VBT User s Guide A VBT Toolbar eje 1 6 3 Zoom Buttons Zoom in and zoom out can be used with the oscilloscope and echogram win dows In addition to the toolbar buttons you may zoom into either window by drawing a box using left mouse button The zoomed in view for the echogram window will appear in a new window Double clicking in the win dow will return the window to its original view 1 6 4 Play Buttons B Play begins data analysis of the data from the beginning of the data file Stop pauses further analysis of the data Ping Navigation Buttons A ping refers to the data obtained from a single pulse sent out by the echosounder A data file consists of hundreds sometimes thousands of pings When the oscilloscope window or any of the method windows is activate the ping navigation portion of the toolbar is also activate Go bad one ping Soto last ping Y Gotospecifio 44 ping number Go first 3 gt gt Advance one ping Use the ping navigation button to move through the pings in the data file in the oscilloscope window or method windows Double clicking on a ping in the echogram window will bring that ping up in the oscilloscope window BS amp E 2005 03 0022 2 2 30 15 Copyright 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Tip the oscilloscope window must be active for the bottom typing buttons to be active VBT User s Guide
7. VBT includes incorporates user input in five types of windows oscilloscope window echogram window output report window map window and bottom typing method windows in order to 1 isolate the bottom echo portion of each ping in the data file and 2 use the bottom echo signal to determine the bottom type PETITS TTE Pype r VET Seabed Elan silir Output Report Window q Dam Trama m i 1041177 42 02 12 00 AP PAD ELN monu 7 87 10411445 02502392 00 AP oe 7500800H T22 37 9808094 7 80 2 1 r11 9 5 UE SELB ki hz 75 08H H 122737 9100000 7 46 2 LR n2 3 37 8Bll kr az 1 0000H bod uum SANANDA 7 2 1 d sn WTF 2 88 2 08 ISNAN F 7 NITEDIUEERCE nei Vn doy This chapter provides detailed explanations for working with each of the VBT windows 02 PRE RELEASE DRAFT Tip Numerical adjustments affect the width of the sampling windows nof their location along the horizontal axis VBT User s Guide 2 Chapter3 VBT Windows 3 1 Oscilloscope Window The Oscilloscope window graphically illustrates a data file by plotting the intensity of each ping along the distance or depth axis When a data file is opened in VBT the default threshold values and sampling windows are not automatically adjusted to the file as such the first step after opening a data file is to adjust these values The threshold values and sampling windows may be adjusted either man ually or numerically
8. You can manipulate the display by manually adjusting the range of the hor izontal and vertical axes or select from linear or log for the axes scale You must select Update to view the changes to axis range The Export button BS amp E 2005 03 0022 2 2 30 59 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter3 VBT Windows allows you to export the coordinates of each feature space into an ASCII file E Restore Move Size Minimize 3 Maxinize x Close 4 Copy CopyTo Report View MultiFile view FCM Algorithm BS amp E 2005 03 0022 2 2 30 Wertical Max amp Min p Type 1 AMO MT DIS pern TAL HEILE 120 DTE ben H arizantal hax amp Min When you right click on the titlebar or window icon lo cated in the titlebar a menu appears Move brings up the double arrow cursor which allows you to click and drag the method window This can also be done by clicking and dragging the window Size allows you to resize the window This can also be done by moving your cursor to any of the corners of the method window and then clicking and dragging Close closes the method window Copy allows you to copy an image of the window to the clipboard which can then be pasted into another program Copy to saves an image of the window to a bitmap file Report View allows you to view the
9. 2 Chapterl Getting Started y Note If you wish to analyze a file more than once Le to see the results of changes made through the CONFIGURATION MENU select the go to first ping but ton and press play again New echogram and output report windows will appear during the reanalysis 1 6 5 Bottom Typing Method Buttons The bottom typing method buttons allow you to select which signal analysis method to use See Chapter 4 2 for a brief description of each method B1 82 83 84 5 Only one bottom classification method can function at a time The data will be classified according to the last open method window or the default method B4 The method in use is highlighted with a red rectangle In depth tutorials for each bottom classification method are included in Chapter 2 B1 First Echo Normalization cumulative curves B2 First Second Bottom Ratio B3 First Echo Division B4 Fractal Dimension S Sediment layer above the bottom 1 6 6 Print amp Help Buttons amp Print prints active window Help opens help file BS amp E 2005 03 0022 2 2 30 16 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Menubar Commands 1 7 Menubar Commands 1 7 1 File Menu After launching the program you can either open a data file or load a pre viously saved configuration file If you open a data file the default configu ration file c windows V VB T INI
10. 56 8 1 7 1990 J C Dezdek Pattern Recognition with Fuzzy Objective Function Algo rithms Plenum Press New York 1981 Z Lubniewski and A Stepnowski Sea bottom typing using fractal dimen sions Proceedings of the International Symposium on Hydroacoustics and Ultrasonics Gdansk Jurata Poland 12 16 May 1997 B B Mandelbrot The fractal geometry of nature Freeman San Fran cisco 1982 A Orlowski Application of multiple echo energy measurements for eval uation of sea bottom type Oceanlogia 19 61 78 1984 E Pouliquen and X Lurton Sea bed identification using echosounder signal European Conference on Underwater Acoustics page 535 A Stepnowski D Bakiera and M Moszynski Analysis and simulation of hydroacoustic methods of sea bed classification Raport Badawczy 52 1995 A Stepnowski D Bakiera M Moszynski and J Burczynski Visual real time bottom typing system vbts and neural network experiment for sea bed classification Proceedings of 9rd European Conference on Un derwater Acoustics Heraklion Crete Greece pages 24 28 June 1996 J Tegowski Characteristic features of back scattering of acoustic signals in south baltic sea ph d thesis charakterystyczne cechy rozpraszania wstecznego sygnalw ultradzwiekowych od dna w baltyku poludniowym doktorat i o pan sopot Ph D thesis 1994 113 PRE RELEASE DRAFT
11. Cumulative Energy Curve 4 2 2 Method B2 First Second Bottom Ratio The First Second Bottom Ratio Method B2 used in VBT is based on a bottom typing method developed by Orlowski 2 and further researched by Chivers et al 1 The method was first commercialized by SonaVision Ltd with their Rox Ann product line Following Chivers et al 1 VBT Method B2 uses the hardness and rough ness signatures of the bottom echo to identify the bottom type EI first part of the first bottom echo contains the hardness signature of the bottom El second part of the first bottom echo contains the roughness signature of the bottom E2 the second bottom echo signal also contains the bottom hardness signature Figure 4 2 2 illustrates the formation of the first bottom echo El and El The initial part of the first bottom echo E1 is caused by the first reflection from the surface perpendicular to the transducer axis This part of an echo specular and coherent is very sensitive to pitch and roll of the vessel and the transducer The remainder of the first bottom echo E1 is caused by oblique back reflection non coherent and is less sensitive to pitch and roll BS amp E 2005 03 0022 2 2 30 6 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Bottom Classification Methods SS os obig P x E p i ma E peculiar M 7 F ur Tn a T E En 7
12. e Fractal dimension Configuration gt Output Report Format Options The Output Report Format Options command in the CONFIGURATION MENU allows you to select the format for the date longitude amp latitude and output file delimiter Date Format C mm dd yy dd mm yy Date Format allows you to specify a date format of pA adem mm dd yy or dd mm yy Latitude Longitude Format Output Fields Separation Symbol Comma Tab BS amp E 2005 03 0022 2 2 30 All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started Latitude Longitude Format allows you to specify a lati tude longitude in NMEA for mat or numerical format See Section 1 7 5 for a description of NMEA and numerical for mats Output Fields Separation Symbol allows you to specify the delim iter between output report columns comma separated value or tab separated value Note Specifications made in Configuration gt Output Report Export Fields and Output Report Format Options will affect both the Output Report window and any exported report file 1 7 4 Export Data Menu Export Data gt Export Pings to File The raw data digital sam ll Export Ping Data ASCII or DT4 BioSonics Inc Em cl x i Sia Gana GRON ples contained in an open NEUES MAPS dt4 file can be exported as a a alle A text file using this command C O
13. of the sea bottom classification with a division of the first echo signal Proceed ings of the XIIIth Symposium on Hy droacoustics Gdynia Jurata 55 60 2 Chivers R C 1994 Acoustical Sea Bed Characterization XIth Symposium on Hydroacoustics Jurata 3 Hastings H M Sugihara G 1994 Fractals A user s guide for the natural sciences Oxford University Press Ox ford New York Tokyo 4 Mandelbrot B B 1982 The fractal geometry of nature Freeman San Fran cisco 5 Pouliquen E Lurton X 1992 Sea bed identification using echosounder sig nal European Conference on Underwa BS amp E 2005 03 0022 2 2 30 106 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Appendix D Factory Settings for Configuration File VBT ini Reports Pings 20 Percent 60 Averaging 1 RectLibrary01 E1Min 60 00000 E1Max 10 E2Min 60 00000 E2Max 20 No 3 Rect1 43 124 111 150 Rect2 96 94 142 141 Info00 5N 4908 0 4 ms BTr 20 log BW 16 48 50 100 16 Info01 MUD120 DT4 red Bot Type 1 120 kHz Rect3 115 52 167 111 Info02 SAND120 DT4 green BT 2 ROCK120 DT4 blue BT 3 Info04 Method 2 E1Max 10 E2Max 20 E1Min 60 E12Min 60 LinLog 1 PointSize 2 Echo window size Rect 0350 0006 0631 0255 icon 0 max 0 Options Echogram Size 171 BiMaxError 0 07 107 PRE RELEASE DRAFT VBT User s Guid apterD iuo s for Configuration File VB T ini bthreshol
14. one with a pulse duration of 7 0 4 ms and the other with a pulse duration of 7 0 4 ms Standard curves for the differ ent pulse durations are displayed in the following figures ChapterA Method 1 Figure A 3 Standard curves acquired for 38 kHz a echoes with maximum amplitude b av eraged for all echoes rock sand soft mud mud amp sand BS amp E 2005 03 0022 2 2 30 86 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide DD Results Figure A 4 Standard curves acquired for 120 kHz a for echoes with maximum ampli tude b averaged for all echoes rock sand soft mud mud amp sand Figure A 5 Standard curves acquired for 420 kHz aJ b echoes with maximum ampli tude c d averaged for all echoes rock sand soft mud mud amp sand BS amp E 2005 03 0022 2 2 30 8 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A 3 Discussion For all processed data the standard curves obtained for maximum amplitudes echo are less distinguishable from each other than in the case of all echoes averaged This indicates that echoes obliquely re flected from the bottom preserves the same or similar shape as echoes reflected perpendicularly Useful results for bottom classification do not have clear dependence from sig nal frequency Standard cumulativ
15. sand gravel and rock for ideal cone beam pattern of a 26 and b 13 width BS amp E 2005 03 0022 2 2 30 91 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 ChapterB Methods B2 amp B3 For a real beam pattern with an obliquely assumed the energy of the echo within reflected pulse with side lobes this part the attack phase boundaries can express of echo arrives when the reverberation the bottom hardness signature surface no longer circular but donut shaped and lasts until the whole pulse penetrates the bottom The limits of appropriate phases can be calculated using some geometric and pulse parameters beam pattern width water depth pulse length etc 2H to CO 2H ti 7 CO 2H to 60 cos 0 2H t3 Co cos 0 where H water depth sound propagation velocity in water sounding pulse length to start of the echo t4 start of the decay phase to start of the release phase t3 end of the release phase 0 half beam width of the transducer Alternatively the phases limits can be found using envelope maximum localiza tion The beginning of decay phase is lo calized at the maximum of an envelope while the end of this phase corresponds to the point where the extension of echo envelope crosses the time coordinate The signal received during the attack phase is primarily due to specular re flecti
16. we can evaluate this dimen sion for such echo pulse not as a limit but for a finite fixed value of r 102 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide C 4 Materials and methods The bottom echoes data we used to cal culate the fractal dimension were recorded in the water region of Lake Washing ton with the use of a digital DT4000 BioSonics echosounder with two operat ing frequencies 38 kHz and 120 kHz Simultaneously the current position of the ship was registered using the GPS system Data acquisition was performed both while the ship was moving along the selected transects and while the ship was anchored In each case the type of sea bottom in the given water region was known This enabled a verification of the applied method of bottom typ ing The length of the pulse sent by the echosounder was 0 4 ms and the sam pling frequency was equal to 41 66 kHz It is not easy to calculate the fractal di mension of a figure following the defini tion of the Hausdorff dimension There fore we decided to use the box dimen sion 3 that can replace the Hausdorff dimension for many sets including shapes of our echo pulses The box dimension for a figure on a plane is defined as fol lows Let N As denote the number of boxes in a grid of the linear scale As which meet the set X on a plane Then X has a box dimension log N As D li rum log As
17. 10 4 2 Bottom tracking sample widths 71 4 3 Echo signal comparison for hard and soft bottoms 76 4 4 Formation of the first bottom echo El and El iri 4 5 Formation of the second bottom echo E2 i 4 6 Succeeding phases of the sounding pulse propagation for an ideal beani Datt fIb v 44 2 4 mex Woo SRE e 18 4 7 Model echo envelope for soft bottom 19 4 8 Model echo envelope for sand gravel and rock 19 4 9 Equilateral triangle containing four self similar equilateral tri cuc awash NE lw eS ee E SO 4 10 A self similar triangle from Figure 4 2 4 B Original triangle in Figure 4 2 4 Magnifying the self similar triangle by two results in the original triangle 0 80 4 11 FCM comparison for different types of bottom 82 A 1 Algorithm of bottom classification with cumulative curves method 85 A 2 Set of normalized and integrated theoretical amplitude envelopes 86 A 3 Standard curves acquired for 38 kHz 86 A 4 Standard curves acquired for 120 kHz 8T A 5 Standard curves acquired for 420 kHz 8T B 1 The results of echo envelope simulations for an ideal beam DAM puede Bw Gand A a A 90 B 2 Succeeding phases of the sounding pulse propagation for an ideal Deam Datter aee ve A Ee E Eee 90 B 3 The results of echo envelope simulation for soft mud 91 vil PRE RELEASE DRAFT viil B 4 B 5
18. 3 1 3 Zooming Zoom in and zoom out are available in the Toolbar and the View Menu In addition you may zoom into the Oscilloscope window drawing a box with the mouse around the area you which to enhance You may zoom out by double clicking in the Oscilloscope window Several double clicks will reset the Oscilloscope window to its original size 3 1 4 Using Color We recommend customizing the colors in your Oscilloscope window for easier interpretation The Color options are available in Configuration gt Color 3 2 Method Windows Although all method windows can be displayed at the same time and running only one bottom classification method can function at a time in VBT The algorithm performs the classification using the criteria established by the last method window opened For an in depth description of working with the method windows please refer to the tutorials in Chapter 2 Working with Method B1 is not included in this manual 3 2 1 Components of the Method Window 3 2 2 Methods B2 B3 amp B4 The layout is the same for the B2 B3 and B4 method windows However The axes and the computational method implemented differ See Chapter 4 2 for details Although the methods described in this subsection apply to all three methods for simplicity the B4 method window is the only window displayed in the figures When working with VBT to classify bottom types you should use cus tomized parameter sets for each bottom type
19. 4 5 8 ss BA AIDISCUSSTON x Vesti mox RU A eS ecu Bio 6000 se amp om Wege cm 6 med C Method 4 CAS AI C2 Mtro 0 4 5 y woe a E A AA Co analysis ut AAA e CA Materials and methods s 200 sii RS es EWOOR TNT TS TTE 0 0 CONCISO e ub ree eR bo es e ee A QUUM S I n 0 2 8 8 oe eS D Factory Settings for Configuration File VBT ini References PRE RELEASE DRAFT 89 39 89 93 97 99 101 101 101 101 103 103 106 106 107 113 vl PRE RELEASE DRAFT CONTENTS List of Figures 1 1 Sample oscilloscope window 4 12 Sample echogram window 5 1 3 Sample output report window 5 1 4 Sample map window 0 e CRI DOE ec IS 6 1 5 The five method windows available in VBT 7 1 6 Sample oscilloscope window 10 1 7 Sample echogram window 12 1 8 Sample output report window 12 3 1 Bottom sampling thresholds in the oscilloscope window 56 3 2 Bottom sampling windows layout 57 3 3 Bottom sampling windows on the oscilloscope 57 ob FCM Algorithm options dicen aaa 6 a 61 4 1 Bottom tracking thresholds
20. 51 88 47 41 6580808H 122736 3200004 1048 8 1 49e 003 4 29e 665 1 88e 889 46 18 18 95 23 29 53 88 47 41 646686N 122736 3200004 1049 8 1 69e 883 1 14e 005 1 88e 889 56 18 18 95 23 29 55 88 47 41 646686N 122 36 320080U 1049 8 1 17e 883 6 40e 006 1 880e 889 66 18 18 95 23 29 57 88 47 41 646666N 122 36 328888U 1034 8 4 18e 004 1 856 888 3 6 76 18 18 95 23 29 59 88 547 51 6580888H 122 36 328888 7 8 2 47e 004 6 40e 006 31 88e 889 86 18 18 95 23 30 01 00 547 51 64580888H 122 36 328888U 2 8 1 47e 004 6 46e 006 1 00e 009 96 18 18 95 23 38 83 88 547 51 658888N 122 36 326666 1656 8 1 28e 885 6 58e 8806 1 66e 669 186 18 18 95 23 38 85 88 547 41 658888N 122 36 328088VU 1051 8 1 61e 004 9 64e 006 1 00e 009 116 18 18 95 23 30 07 00 47 41 658888N 122 36 3280088VU 1656 8 5 26e 004 1 04e 005 1 88e 889 126 18 18 95 23 38 89 88 47 51 658888N 122 36 328888U 1052 8 2 38e 004 6 46e 666 1 00e 009 16 16 95 646666N 122736 8 4 85e 004 6 286 886 1 00e 009 Figure 1 8 Report Window for file sand120 dt4 You can choose to exclude include data fields by going to Configuration lt Output Report Export Field You can alter the date and direction format by going to Configuration gt Output Report Format Options Step 4 Interpreting the Bottom Typing Results Step 4a Echogram window during analysis the color code See Sec tion 3 3 of the bottom type for each ping will appear along the bottom axis BS amp E 2005 03 0022 2 2 30 12 Copyright
21. 60 000 Fo 90000 Info Step 10 Right click on the method window titlebar and select Report View lhis will display only the pings for the rock data file that passed through the filter All pings will be iic in red Fractal Dimension B4 Method BroSonics nc R DOCE 60 000 Step 11 Right click on the method window titlebar and select Multifile View This will display all pings that passed through the filter for all three BS amp E 2005 03 0022 2 2 30 40 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Ground truth Library files The pings for mud will be red the pings for sand will be green and the pings for rock will be blue Fractal Dimension B4 Method S BioSonics Ir x PL peu Inu Info gt gt ED import Export Step 12 Right click on the titlebar and select FCM Algorithm Fuzzy C Means Clustering BioSonics Inc E x Number of clusters Database option Cancel E Display cluster centers only Append to current database Rectangle size 4 Horizontal axis multiplier Vertical axis multiplier References Bezdek J C 1881 Pattern Recognition Algorithms with Fuzzy Objective Functions Plenum Press New York Step 12 a Select Create new database in order to add feature space box
22. EVEN IF BIOSONICS INC HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGE SOME JU RISDICTIONS MAY NOT ALLOW THE EXCLUSION OR LIMITATION OF LIABILITY FOR CONSEQUENTIAL OR INCIDENTAL DAMAGE BioSonics Inc s entire liability and your exclusive remedy as to the media shall be at BioSonics Inc s option either a return of the purchase price or b replacement of the defective media Any replacement media will be warranted for the remainder of the original warranty period or thirty 30 days whichever is longer In no event shall BioSonics Inc s liability arising under any cause of action exceed the purchase price of the Software paid by the Licensee PRE RELEASE DRAFT Miscellaneous This Agreement shall be governed by the laws of the State of Washington If for any reason a court of competent jurisdiction finds any provision of this Agreement or portion thereof to be unenforceable that provision of the Agreement shall be enforced to the maximum extent possible so as to effect the intent of the parties and the remainder of this Agreement shall continue in full force and effect PS Ex PRE RELEASE DRAFT PRE RELEASE DRAFT Installation Instructions 1 Insert BioSonics Inc VBT SeaBed Classifier installation disk into your CD ROM drive If the installation program does not be gin automatically browse to your CD ROM drive and double click on setup exe The following window will appear 2 Select Next
23. Inc All Rights Reserved PRE RELEASE DRAFT VBI User s Guide BioSonics Menubar Commands rock120 DT4 anl x Ch 1 Ping 7 Gridines Scales Ticks Goto Ping Number ey Gobo First Ping i Grid displays gridlines in the oscilloscope window Thresholds displays a dotted horizontal line for the Bottom De tection Threshold and the Peak Threshold defined in Configuration gt Bottom Tracking Parameters Bottom Lines sampling windows for bottom tracking defined in Configuration lt Bottom Tracking Parameters Scales Ticks display axes labels and tick marks in the oscilloscope window View gt Echogram The echogram submenu allows you to select or deselect the axes to view 500 640 S80 720 760 800 890 880 320 Aut Ping Number Echo Intensity Color Bar dB Bottom Classification Color Bar Ping Scale ping number is the horizontal axis in the echogram window Depth Scale the depth of each sample in each ping is plotted ver tically Echo Level Color Bar the relationship between the color of each sample and its energy intensity BS amp E 2005 03 0022 2 2 30 19 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started Bottom Classification Color Bar the color code of the bottom type for each ping Bottom Trace draws a vertical line indicating the presence of a
24. Jim C 2 The method of evaluating the box di mension of the bottom echo envelope is explained in Fig 2 We evaluated the box dimension D for 2 EDITOR S NOTE for simplicity the authors term box dimension may be considered equivalent to the fractal dimension CJC BS amp E 2005 03 0022 2 2 30 D BioSonics Results ILI Panna 8E i Bae E PE pump AREER E tT LE Figure 62 2 Illustration of the box dimension evaluation In the considered case As 0 1 N As 30 the finite value of As 1 36 that is there were 36 boxes both horizontally and vertically on the area of the echo waveform plot Each echo pulse was nor malized to the standard length and height and afterward the number of boxes N As was counted as shown in Fig 2 The box dimension was evaluated according to formula 2 without using the limit The amplitude threshold used for the analyzed signals was 70 dB The box dimension was calculated for each echo pulse separately and histograms of its values for each type of bottom were con structed and analyzed C 5 Results Figures 3 4 and 5 show the histograms of box dimension values obtained for echo pulses for four types of bottom Three sets of acquired data were analyzed sta tionary data of echosounder frequency 120 kHz and 38 kHz and data from tran sects data of echosounder frequency 120 kHz
25. PRE RELEASE DRAFT VBT User s Guide 2 ChapterB Methods B2 amp B3 An additional set of 420 kHz data was AAA analyzed in order to find the relationship between bottom hardness and roughness signatures and the grain size of bottom sediments The percent of particles of different grain size is given in the table below files PEPIN from Bruce Sable Core samples of the upper part of the bottom 10 cm thick were taken 0 002 0 04 0 06 8 0 1 El measure ofroughness Table B 1 Percentage contents pina grabediattiste of particles of different size in bottom samples of of of tt particles particles particles E gt 50 um 2 um gt 2 um E gt gt E 50 um 2 i 1 E 2 1 3 4 5 D Off 01 015 02 025 0 035 04 6 El measure afroughness x 1m cdna dmbroziamiste 35 9 10 11 as mainly mussels and fragmented mussels percentage 5 2 contents is not valid for this sample E i 15 a 5 0 0 0 04 0 06 D ns 0 1 El measure afraugness didra gruboziami ste D O08 014 015 OF 025 Of OS 04 El measure ofroughness Figure Classification results for bottom of differ ent grain size frequency 420 kHz pulses with maximum am plitude a b first echo divi sion method c d second echo BS amp E 2005 03 0022 2 2 30 96 Copyriot Ha 998 2065 2BWH80n4c5 Inc 1 x 5 6 AlI Right Reserved 10_ b d coarse grain 2 x 3 4 0 9 11 PRE RELEASE DRAFT VBT Us
26. Peak Width 53 ram b IE Bottom Detection Threshold dB Above Bottom Blanking Zone samp como camp Pulse Width zamp Alarm Limit Tracking window Cancel Tracking Domain 20logR 40logR BS amp E 2005 03 0022 2 2 30 1 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter4 How VBT Works Bottom Tracking Parameters Peak Threshold The minimum intensity an echo must reach for PEAK WIDTH number of samples to be considered a bottom echo Peak Width The minimum number of consecutive samples an echo must be above the PEAK THRESHOLD to be considered a bottom echo Bottom Detection Threshold The maximum intensity a portion of the echo must fall belov ABOVE BOTTOM BLANKING ZONE number of samples to be considered an endpoint of the bottom echo Above Bottom Blanking Zone The minimum number of consec utive samples an echo must be below the BOTTOM DETECTION THRESHOLD to be considered a bottom echo Alarm Limit The number of consecutive lost bottoms before the bottom tracking algorithm is reinitialized and the bottom echo is tracked from the bottom of the last ping s bottom echo Bottom Tracking Window Search window for the bottom tracking algorithm The window is centered on the bottom line of the last found bottom and the algorithm searches for the next bottom within the window in both directions Lost Bo
27. RELEASE DRAFT VBT User s Guide 2 Chapter4 How VBT Works echo decaying part of the pulse after signal maximum contains the bottom roughness signature The bottom echo is formed in three distinct phases Figure 4 2 3 Phase 1 attack from the moment the pulse reaches the bottom until the time when the bottom is reached by the back slope of the pulse Phase 2 decay beginning at the end of attack phase and lasting until the time when the front of the pulse reaches the boundary of the ideal beam pattern Phase 3 release lasting until the time when the pulse completely enters the bottom During Phase 1 the echo is due to a surface reverberation the reverberation area is circular with diameter increasing with time time equal approximately to pulse duration Phase 2 is combination of surface and volume reverbera tion During Phase 3 the volume reverberation determines the bottom echo and the reverberation area has a shape of doughnut Pana Plans ene Pleas ECT E E i i X Figure 4 6 Succeeding phases of the sounding pulse propagation for an ideal beam pattern After the release phase is complete in the case of an ideal beam pattern the only source of the echo can be the reverberation due to bottom volume inhomogeneity This can be observed especially in the case of soft bottom types for which the main part of the echo energy penetrates the bottom The example i
28. Select Bottom Sampling Windows from the CONFIGURATION MENU The data acquisition pulse is included in the window for reference and cannot be altered Table 2 1 Recommended Bottom Sampling Window Widths Sampling Window Methods B2 B3 amp B4 E E 1 3 x pulse duration 2 6 x pulse duration 6 x pulse duration S 50 samples 50 samples Bottom Sampling Windows BioSonics Inc E1 First Bottom First Part samp El First Bottom Second Part jaa samp Cancel E2 Second Bottom Window samp Sediment Window samp Data Acquisition Pulse Duration 6 samp BS amp E 2005 03 0022 2 2 30 39 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter2 Tutorials Step 5 Go to Configuration gt Output Report Filters Deselect stop analysis after one report and leave all other parameters in their de fault settings Energy filter E Pings per Report Cancel iv Report the average of all qualifying pings Stop analysis after one report Step 6 Begin analysis of the data file by selecting play Step 7 Open sand120 dt4 verify stop analysis after one report is deselected and analyze the data file Step 8 Open rock120 dt4 verify stop analysis after one report is deselected and analyze the data file Step 9 Open the B4 Method Window Fractal Dimension 64 Method BioSonics In EE Dont
29. according to the last open method window or the method selected via Configuration gt Select Bottom Typing Method Step 2e Select the ground truth data set to use for bottom typing If you have not yet created your own ground truth data set use the data set included with VBT ID 01 in the method window Step 3 Data Analysis Bottom Tracking amp Bottom Typing Step 3a There are three ways to begin analysis of the data i Press the play button located in the toolbar BS amp E 2005 03 0022 2 2 30 11 Copyright 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Bottom Typing B1 82 83 fea 5 Play VBT User s Guide A Chapterl Getting Started ii Select Play from the ANALYZE MENU ii Press FO Step 3b When analysis begins an echogram window and output re ports window appear on the screen BP SANDI70 0T4 3 E JO ail Figure 1 7 Echogram window for file sand120 dt4 See Section 3 3 for more information on interpreting echogram windows PA ing amp Echo Level Colorbar dE 1500 2000 SAND120 DT4 2 IN aloj xl Date Time Latitude Longitude Depth Type 0 2 18 18 95 23 29 45 47 41 122736 3200004 1655 8 1 44e 003 1 16e 005 1 88e 889 16 18 18 95 23 29 27 88 547 51 658888N 122 36 328080VU 1050 8 8 13e 004 6 46e 666 1 00e 009 26 18 18 95 23 29 49 00 547 41 658888N 122736 3200004 1051 8 3 46e 004 6 40e 006 1 00e 009 36 18 18 95 23 29
30. data file is open you can Open a Data File allows you to browse through your directory to select a data file Close a Data File closes the open data file and all as sociated windows Export The Ping Data To A File allows you to ex port pings to a file in either ASCII format or in BioSonics Inc DT4 format In ASCII format each line of the text file is a single ping and each number contained in the line is a sample Refer to Section 1 7 4 for more information on this com mand Print allows you to print the active Oscilloscope window Recent File List allows you to select from the last four data files opened in software 1 7 2 View Menu YJE w Toolbar w Status Bar Oscilloscope Echagram POCHI AO ute Ping Number err SOLON ars EMU ss anta Previous Pings Soto MextP Ing Goto East Ping gt gt Shaw Map Window Zoom In scope Echogram Map zaon out scapge Echodram Mapi As described in the following subsections the VIEW MENU allows you to GEFI CE TEL 7 1 customize the oscilloscope window crite 2 customize the echogram window zu 3 navigate through the pings in the data file Ctrl 4 zoom in the oscilloscope echogram or map windows 5 show the map window View gt Oscilloscope In the Oscilloscope submenu you may select whether or not to display the following objects in the Oscilloscope window BS amp E 2005 03 0022 2 2 30 18 Copyright C 1998 2005 BioSonics
31. manner in which sample windows are de fined in the Oscilloscope window Please note the shared boundaries between the S El and El windows Intensity Axis dE Distance mj Start 082 Echo Twice Start of 1 Echo the distance of the start ofthe 17 Echo Figure 3 2 An illustration of the arrangement of sampling windows in the Oscilloscope window Please note the shared boundaries between the S El and El windows In the Oscilloscope window the sampling windows are represented by verti cal dotted lines You can manually adjust the width of the sampling windows by dragging right hand boundary of the window Dragging the left hand boundary of the window allows the user to shift both sides of the window simultaneously while maintaining the width of the window Clicking on the left hand boundary of El the start of the bottom echo will allow you to shift all windows simultaneously while maintaining their width and relative positions The sampling window boundaries in the Oscilloscope window for 5 El and End af E1 Sample window End af E1 and start of El 3L 1 Figure 3 3 An illustration of the sampling window boundaries in the Oscil loscope window for the sediment window and first bottom echo both parts BS amp E 2005 03 0022 2 2 30 57 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Zoom in amp Out VBT User s Guide 2 Chapter3 VBT Windows
32. signal using any of four peer reviewed algorithms as outlined in Table 1 1 The complete references for the classification methods are provided in the Bibli ography on Page 113 This user s guide demonstrates how to use BioSonics Inc VBT SeaBed Classifier to classify the bottom structures present in your hydroacoustic data Chapter 1 contains an introductory tutorial for VBT as well as de scriptions of the toolbar and menubar options Chapter 2 includes in depth tutorials for creating a ground truth library as well as analyzing a data file Chapter 3 describes how to manipulate each of the windows used in VBT Chapter 4 provides a brief description of the algorithms included in VBT work It is not necessary to read the guide from cover to cover In order to main tain chapter independence we have repeated important points throughout the text Information that is only provided once is cross referenced in related Table 1 1 Summary of Bottom Classification Methods Method Echo Signal Feature Means of Comparison IET WE Cumulative Energy sd the Least Squared Error BN o o NN Energy Ratio E1 E2 Cluster Analysis Second Part of Bottom Echo and the Second Bottom Echo 3 7 Energy Ratio E1 E1 of the Cluster Analysis First and Second Parts of the First Bottom Echo 2 B4 Ratio of the Fractal Dimension Fractal Dimension of the envelope of the first amp bottom echo energy E1 FD
33. signal normalized to the total energy ac cording to the following formula s r dr gm s t dr where s 7 is the amplitude of echo en velope signal E t E pe 0 Tmar The echo signal envelope is digitized and than processed In order to minimize the fluctuation of the echo envelope the normalized sum is calculated as a sig nal integral in a discrete time As a next step of processing the obtained re sult is averaged over a number of pulses e g 5 to 10 Averaged cumulative echo curves are compared with theoreti cally obtained model curves correspond ing to different bottom types The curve most similar to the received echo de termined using minimum mean square error algorithm is treated as the corre sponding bottom type The schematic 9 PRE RE diagram of the processing algorithm is shown on Fig 1 PROCESSING Echo envelope from Echo Sounder e Analog to Digital Conversion Bottom Detection Signal Packagin Integration Normalization dl Averaging Comparaison Identificatian Figure A 1 Algorithm of bot tom classification with cumula tive curves method ih Pouliquen and Lourton used seven stan dard curves for bottom type identifica tion which were obtained using model of bottom reverberation for rock gravel sand very fine sand muddy sand mud and soft mud Reverberation model is based on Kirchoff method tangential plane ar
34. the capabilities of VBT using the sample library before integrating your own data After you practice with the sample library your own verified library can be integrated by first taking physical samples of the bottom and simultaneously recording the bottom echo signals Once the data has been collected you can create and save customized feature spaces for each bottom types using the 37 PRE RELEASE DRAFT VBT User s Guide 2 Chapter2 Tutorials method windows The customized feature space created from the physically verified data sets is referred to as ground truth library The ground truth data is used to classify future data sets Verified standard databases for different types of the bottom should be stored in VBT using the bottom typing method windows B1 B2 B3 B4 S Echo signal parameters depend not only of bottom type but also on data acquisition parameters i e transducer frequency beam width pulse length etc Therefore the ground truth data set is valid for a particular set of equipment and for particular equipment parameters 21 Ground truth Library The purpose of this tutorial is to illustrate 1 creating the ground truth library and 2 classifying a transect according to the ground truth library There are two methods for creating a ground truth library in VBT 1 automatically generating the feature space boxes using the Fuzzy C Means FCM clustering algorithm and 2 manually generating t
35. them are not in line with the expectation 104 Copyright 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT D VBT User s Guide Percentage of pings 3s zb 2 T 10 5 Figure C 4 The histogram of the box dimension values evalu ated for echo pulses from station ary data for the echosounder fre quency 38 kHz a for all echo pulses b for selected 10 of echo pulses of the highest ampli tude level that echo envelopes from a bottom with higher hardness should have a higher value of the fractal dimension This may be due to the fact that when pulses at this frequency are reflected from the bottom for certain reasons the fractal structure of the bottom may fail to transfer itself onto its image in the echo envelope BS amp E 2005 03 0022 2 2 30 BioSonics Results Percentage of pi regs SO 25 7 1556 10 Box dmereion Figure C 5 The histogram of the box dimension values evalu ated for all echo pulses acquired from transects for four types of bottom at the echosounder fre quency 120 kHz a the echo pulse from the first class b the echo pulse from the second class Figure C 8 Sample oscillo grams of two classes of echo pulses recorded for the rocky bottom from transect 105 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide C 6 Conclusion The results of the pre
36. to bring up an electronic copy of the user s guide BS amp E 2005 03 0022 2 2 30 35 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started BS amp E 2005 03 0022 2 2 30 36 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Chapter 2 Tutorials As described in Section 1 5 there are five major steps for analyzing a data file using VBT loading the data file preparing the data file for analysis analyzing the data file interpreting the analysis results and exporting the bottom typing results Before a data file can be analyzed a ground truth library must be created by the user This section describes 1 how to create a ground truth library and 2 how to analyze the data file using the ground truth library BioSonics Inc VBT SeaBed Classifier includes five methods of bottom typing B1 First Echo Normalization cumulative curves B2 First Second Bottom Ratio B3 First Echo Division B4 Fractal Dimension S Sediment layer above the bottom y Note Although all method windows can be displayed at the same time the algorithm will perform classification of the bottom according to criteria of last open method window VBT also includes a practice library found in the CTLG folder with data files for various bottom types sand rock mud and soft mud We recom mend you familiarize yourself with
37. will be used y Note The default configuration file will be updated after VBT or a data file is closed We recommend backing up VBT ini found in C Windows before beginning to use VBT This will allow you to restore the default settings in the future Before a data file is open Before a data file is open you have the option to aE Open Data File Ckri O Load Configuration File Load Configuration File choose this Op save Configuration File tion if you wish to use a Configuration file saved Setup previously The settings saved in the previously 1 Tls sanpizo o74 i l 2 CTLGITR _B DT4 saved Configuration file will be applied during 3 CTLGIMUD120 DT4 4 CTLGIROCK 120 074 this session Pe ExiE AlE F4 Save Configuration Filet Open a Data File allows you to browse through your directory to select a data file If you do not need to load or save a Configuration file simply open a data file y Note Once you option a data file you do not have the option to change the Configuration file BS amp E 2005 03 0022 2 2 30 17 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started After a data file is open pe Wwe Corea Oper Daba 8 do Cola Ale Esport Am ez Fue Fori Prii Era mem bnt sema o 2 BOTA Ace oe Ed a tz Che 4 Chie Chi Hii 1 a When a
38. 0 Info00 120 kHz 0 4 ms SN 49408 BotWin 16 48 PL 16 Info01 M0D120 DT4 red Bot Type 1 Info02 SAND120 DT4 green T2 ROCK120 DT4 blue T3 VBT Files C BIOSONICS VBT CTLG ROCK120 DT4 43 69 50 16 48 100 0 23870 55 9157 5 222 58 0 4 120000 1490 34 0 6 0 000990972 70 70 49408 C BIOSONICS VBT CTLG SAND120 DT4 48 69 50 16 48 100 0 23870 55 9157 5 222 58 0 4 120000 1490 34 0 6 0 000990972 70 70 49408 C BIOSONICS VBT CTLG MUD120 DT4 50 69 50 16 48 100 0 55 9157 5 222 58 0 4 120000 1490 34 0 6 0 000990972 70 70 49408 C NBIOSONICSNBTNCTLGMMUDA20B DT4 2 55 0047 69 50 16 48 100 0 23870 56 0601 2 218 51 0 2 420000 1486 5 C BIOSONICS VBT CTLG SAMD420B DT4 55 69 50 16 48 100 0 23870 56 0601 2 218 51 0 2 420000 1486 5 BS amp E 2005 03 0022 2 2 30 109 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guid apterD for Configuration File VB T ini C BIOSONICS VBT CTLG MUD420 DT4 63 69 50 O 24 100 0 23870 56 0601 2 218 51 0 2 420000 1486 5 C BIOSONICS VBT CTLG SAND420 DT4 55 69 50 O 24 100 0 23870 56 0601 2 218 51 0 2 420000 1486 5 C BIOSONICS VBT CTLG MUD120 DT4 48 69 50 16 48 100 0 23870 55 9157 o 222 58 0 4 120000 1490 34 C BIOSONICS VBT CTLG TR7_B DT4 54 69 50 16 48 100 0 23870 55 9157 1 222 58 0 4 120000 1490 34 O 6 0 000990972 70 70 49408 C NBIOSONICSNVBTNCTLGNROCK120 DT4 2 40 69 50 16 48 100 0 23870 55 9157
39. 0 420 kHz SN49522 6 deg PL 0 2 ms Info01 B Windows 0 24 50 100 8 InfoO2 red 1 MUD420 DT4 green 2 SAND420 DT4 Ping20 0 0003 0 0029 0 0178 0 0613 0 1405 0 2473 0 3718 0 5098 0 6531 0 7789 Ping21 0 8640 0 9087 0 9330 0 9518 0 9688 0 9814 0 9883 0 9919 0 9946 0 9971 Ping22 0 9988 0 9996 0 9999 1 0000 Method 4 XMax 1 10001 YMax 0 XMin 0 9 YMin 60 LinLog 1 View Map 0 Map window size Rect 0101 0024 0591 0414 icon 0 max 0 Map BS amp E 2005 03 0022 2 2 30 111 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guid apterD for Configuration File VB T ini Zoom 1 Map Files TR7_B DT4 map mapa bmp 47 7177 122 637 47 6997 122 596 1 11 0 MUD120 DT4 maplemptymap 0 0 1 0 HASP Key 1972 install user Generic company Generic serial number 144 4490 204 just installed 0 BS amp E 2005 03 0022 2 2 30 112 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Bibliography 1 D Bakiera Numerical analysis of acoustic echo signal from the bottom 2 110 11 Lo LL thesis 92 1995 D Bakiera and A Stepnowski Method of the sea bottom classification with a division of the first echo signal Proceedings of the XIIth Sym posium on Hydroacoustics Gdynia Jurata pages 55 60 R C Chivers N Emerson and D Burns New acoustic processing for underway surveying The Hydrographic Journal
40. 0022 2 2 30 64 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Output Report Window EP P dx where EP is the energy of the it ping and P is the ping plotted against depth x EP is calculated for all pings in the data file and the highest value of EP called EP is determined The energy of each ping is compared to E P using the ENERGY FILTER an output report line is created for all EP gt x 121 where x ENERGY FILTER For example the energy filter shown in the figure to the left allows samples with a value of 60 or more EPma to pass and be analyzed by the VBT algorithm The Pings per Report option Bl Configure Analysis Output Report BioSonics Inc AX allows you to specify the step E filter 2 size between pings displayed in Fings per Heport Cancel the output report window as i Report the average of all qualifying pings well any exported report fie When the option Report the average of all qualifying pings is selected the algorithm re turns the average all pings that pass through the filters in a ping step For example the output for ping 31 when pings per report is 20 is the average of all values over 60 for pings 12 31 When the averaging option is not selected all pings that pass through the filter are displayed Stop analysis after one report The option Stop analysis after 1 r
41. 2 2 30 63 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter3 VBT Windows 3 4 Output Report Window The output report window is organized by ping number and displays the data numerically for the following fields Date date the data was collected Time time the ping was recorded Latitude latitude of the ping transmission Longitude longitude of the ping transmission Depth depth to the bottom Type numerical code corresponding to the type of bottom O no bottom coded 1 corresponds red parameter box first file opened 2 corresponds to the green parameter box second file opened 3 corresponds to the blue parameter box third file opened EO energy of sediments echo El energy of second part of 1st bottom echo E2 energy 2nd bottom echo E1 energy of first part of 1st bottom echo Sediment thickness of the sediment layer FD Fractal dimension In the Configuration menu Output Report Filters allows you to select the energy filter and the number of pings per report Each line in the Output Re port Window is one report The value entered in the energy filter specifies the lowest amount of energy a ping can have to be included in the analysis The energy filter is in terms of the percent of the maximum energy present in the data file and can be described using the following mathematical ap proximations BS amp E 2005 03
42. B 6 IT B 8 B 9 C 1 C 2 C 3 C 4 C 5 C 8 LIST OF FIGURES The results of echo envelope simulations for bottom types 91 Classification results for first echo division method 94 Classification results for the second echo method 94 Classification results for the second echo method 95 Classification results for bottom of different grain size 96 Classification results for bottom of different grain size 98 The Koch snowflake 2L 0 6 boe 2 X9 bee dw Rex 4 102 Illustration of the box dimension evaluation 103 The histogram of the box dimension values 104 The histogram of the box dimension values 105 lhe histogram of the box dimension values evaluated for all echo pulses acquired from transects for four types of bottom at the echosounder frequency 120kHz 105 Sample oscillograms of two classes of echo pulses recorded for the rocky bottom from transect 0 105 PRE RELEASE DRAFT Vel 1 2 1 3 1 4 Zl 212 4 4 2 B 1 List of Tables Summary of Bottom Classification Methods 1 Sample library included with VBT 0 00 8 Recommended Bottom Sampling Window Widths 25 Import Export File Formats 32 Recommended Bottom Sampling Window Widths 39 Recommended Bottom Sampling Window Widths A4 Guidelines for Se
43. Buttons 5 6 244896 x ri do a a 15 1 6 5 Bottom Typing Method Buttons 16 1 6 6 Print amp Help Buttons 0 20 16 L7 Menubar Commands eg 2 0 s 3o X Rede EOxs 17 Sr MEE UM e A 17 l2 View Mont sapa ra a A A 18 1 7 3 Configuration Menu 21 1 7 4 Export Data denis aa 0 a x4 90 Lro Edt Map Menis aca ee TES EE GH 33 1350 Analyze NOB e sos daa ra fex 34 1 7 7 Window yo ood dor X Se KR oe Bw de SG 35 LeS Melo Menis eade uu 2 bd Ade cdo dede os 30 2 Tutorials 37 2L Ground LIDI avia 9 ee ee N 38 2 1 1 Creating a Ground truth library using Fuzzy C Means ECONO C USTI ci Roe Oe e e nre 38 2 1 2 Manually generating a ground truth library 43 2 1 5 Classifying data using your ground truth library 48 2 2 Preparing Your Data File for Analysis 49 2 2 1 Oscilloscope Window 49 2 2 2 Echogram and Output Report Windows Settings 50 2 2 3 Bottom Typing Method 50 Zo Analyzing Your Data File o ssia Eo xd eed 51 2 3 1 Single Transducer Data Files ol 111 PRE RELEASE DRAFT 2 3 2 Analyzing Multiplex Data Files 2 4 Interpreting the Analysis Results 2 4 1 Echogram Window 2 4 2 Output Report Window 24 3 Wap Window diia SE ys 2 4 4 Impr
44. C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Basic Steps Step 4b Output report window during analysis the numerical code See Section 3 4 of the bottom type for each ping will ap pear in the Type column Step 4c Map window during analysis if the map window is open the color code of the bottom type for each ping will appear at the corresponding coordinates in the map window Step 5 Export the Bottom Typing Results See Section 1 7 4 for information on exporting the bottom typing results BS amp E 2005 03 0022 2 2 30 13 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started 16 VBT Toolbar 1 6 1 File Buttons Open data Close data file Create new i i M configuration file Tes ds m 4 6 Create New Configuration File resets to factory settings Open Data File allows user to browse directory in order to select data file Close Data File closes data file and all associated windows Save inactive use the export data commands in the menubar 1 6 2 Edit Buttons Cut inactive Copy allows you to copy objects to the clipboard for pasting the object in other programs Paste allows you to paste a bitmap image from the clipboard into the Map Window 14 Copyright C 1998 2005 BioSonics Inc All Rights Reserved BS amp E 2005 03 0022 2 2 30 PRE RELEASE DRAFT
45. In each case more than 600 pings for each bottom type were taken for the calculations For stationary data apart from analyzing the distribution of values of box dimension for all pulses 10 of pings of the highest amplitude were se 103 Copyright 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide lected in order to take into account only the echoes from pulses with the most likely normal incidence to the bottom and to drop the others reducing the ef fect of a ship s pitching and rolling How ever it is easy to observe that this op eration had no significant impact on the results The presented histograms show that in the case of the 120 kHz frequency of the echosounder Fig 3 there is a clear dif ference between the values of the box dimension for a rocky bottom and for other types of sea bottom especially for the data collected from the anchored ship In the case of rock the box dimension values are much higher which is in line with expectations because the surface of a rocky bottom is more corrugated and irregular For the other types of bottom the obtained box dimension val ues at a 120 kHz echosounder frequency do not allow for a clear distinction be tween them However in the case of the data obtained from transects the gen eral regularity is evident namely that the harder the bottom is the higher the box dimension values What is interesting is th
46. LEASE DRAFT VBT User s Guide 2 Echogram Window 3 3 Echogram Window lhe echogram window displays the results of the analysis by graphing the ping number top horizontal axis against bottom depth vertical axis The depth of each sample in each ping is plotted vertically xm Ping Number Echa Intensity Color Bar dE Bottorn Classification Color Bar The echo intensity is reflected in the color of each pixel of the Echogram The echo intensity color bar is dis E played on the right hand vertical Step _ Cancel axis The color bar can be ad justed by going to Configuration gt Echogram Color Bar and enter ing the Minimum Intensity dB Maximum Intensity dB and step between intensities dB The order of the colors cannot be manipulated Minimum value 70 The bottom classification of each ping is depicted along the bottom axis The colors correspond to the colors used in the active Method window Red corresponds to the first file opened while creating the feature space green corresponds to the second file and blue corresponds to the third file You can zoom into the Echogram window 1 by using the zoom in and zoom out buttons in the toolbar 2 by selecting the options in the VIEW MENU or 3 by pressing or respectively Boxing a portion of the Echogram window will open a new Echogram window using the coordinates of the box BS amp E 2005 03 0022
47. Line Samples y Note 20logR is the proper setting for the theoretical model of a non point source reflector Beam Type Channel used for multiplex data files where the D T X system used more than one transducer to specify which channel to analyze y Note Only one channel can be analyzed at a time Depth X Scale You may view the bottom depth in units of digital samples or meters You may manually set the range of the horizontal axis by entering values in the Min and Max fields Amplitude Scale You may view the intensity of the signal on a logarith mic or linear scale You may customize the range of the vertical axis by entering values in the Min and Max fields Background Color gt This function does not work properly for this release of VBT 22 Copyright C 1998 2005 BioSonics Inc All Rights Reserved BS amp E 2005 03 0022 2 2 30 PRE RELEASE DRAFT D VBT User s Guide BioSonics Menubar Commands fixed solid color sets the background of the Oscilloscope window to the color defined in Configuration lt Colors gt Scope Echogram Colors sets the background of the Oscilloscope win dow to coincide with the Echogram intensity colorbar defined in Configuration lt Echogram Color Bar along the intensity axis vertical axis Chart Type See Section 3 1 for additional information on working with the Oscilloscope window Configuration lt Echo
48. N consecutive samples then drops below a surface threshold for M samples Once a bottom echo has been identified a bottom sampling window is used to find the next echo The bottom echo is first isolated by user defined threshold values that indi cate 1 the lowest energy to include in the bottom echo bottom detection threshold and 2 the lowest energy to start looking for a bottom peak peak threshold The bottom detection threshold allows the user to filter out noise caused by a low data acquisition threshold The peak threshold prevents the algorithm from identifying the small energy echoes due to fish sediment or plant life as a bottom echo 4 Peak Threshold Intensity 2215 dB H Y Bottom Detection Threshold samples Axs Figure 4 1 Bottom Tracking Thresholds The Bottom Detection Threshold eliminates noise due to the data acquisition threshold from the bottom echo The Peak Threshold prevents the algorithm from classifying small echoes such as fish echoes as bottom echoes To avoid incorrectly identifying small dips or small jumps in the signal as the end of the bottom echo or the peak of the bottom echo the user can define thresholds windows The threshold window indicates the smallest number of BS amp E 2005 03 0022 2 2 30 0 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT D BioSonics VBT User s Guide bottom Tracking samples in a row that must exceed th
49. Paste Map Image from Clipboard In order to paste a map image from a file the Show Map Window command must first be selected VBT supports map images of the following formats bitmap bmp Windows Metafile wmp and Graphic Interchange For mat gif If a supported image file is in the clipboard you may paste it into an active map window using this command This can also be achieved by selecting the paste button in the toolbar y Note VBT is indiscriminate with respect to map files If you import any supported graphics file VBT will assume the file is a map of appropriate coordinates and plot the results accordingly BS amp E 2005 03 0022 2 2 30 33 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Play Go to 196 Ping Stop al VBT User s Guide 2 Chapterl Getting Started Edit Map lt Calibrate Map Spatial Range When a map window is active you may specify the coordinates of the map file using this command window You must specify the coordinates of any map file that was not generated by BioSonics Inc Otherwise the bottom types will not be displayed in the map window Longitude 122738220000w A 47 41 982000N The spatial range of the map is established by entering the latitude and longitude of the top left and lower right hand corners of the map image in degrees decimal minutes Decimal minutes are converted to minutes seconds by multiply
50. TNESS FOR A PARTICULAR PURPOSE AND MERCHANTIBILITY However BioSonics Inc warrants to the original Licensee that the media on which the Software is recorded is free from defects in materials and work manship under normal use and service for a period of ninety 90 days from the date of delivery as evidenced by a copy of the receipt of purchase Fur ther BioSonics Inc hereby limits the duration of any implied warranty ies on the media to the period stated above Some jurisdictions may not allow limitations on duration of an implied warranty so the above limitation may not apply to you THE ABOVE ARE THE ONLY WARRANTIES OF ANY KIND EITHER EXPRESS OR IMPLIED THAT ARE MADE BY BIOSONICS INC ON THE SOFTWARE NO ORAL OR WRITTEN INFORMATION OR AD VICE GIVEN BY BIOSONICS INC ITS RESELLERS DISTRIBUTORS AGENTS OR EMPLOYEES SHALL CREATE A WARRANTY OR IN ANY WAY INCREASE THE SCOPE OF THIS WARRANTY AND YOU MAY NOT RELY UPON SUCH INFORMATION OR ADVICE THIS WAR RANTY GIVES YOU SPECIFIC LEGAL RIGHTS YOU MAY HAVE OTHER RIGHTS WHICH VARY ACCORDING TO JURISDICTION Limitation of Liability NEITHER BIOSONICS INC NOR ANYONE ELSE WHO HAS BEEN INVOLVED IN THE CREATION PRODUCTION OR DELIVERY OF THE SOFTWARE SHALL BE LIABLE FOR ANY DIRECT INDIRECT CONSEQUENTIAL OR INCIDENTAL DAMAGE INCLUDING DAM AGE FOR LOSS OF BUSINESS PROFIT BUSINESS INTERRUPTION LOSS OF DATA AND THE LIKE ARISING OUT OF THE USE OF OR INABILITY TO USE THE SOFTWARE
51. The values can be defined numerically by going to Configuration lt Bottom Tracking Parameters for threshold values and Configuration gt Bottom Sampling Windows for the sampling windows Adjusting the values numerically adjusts the visual display in the Oscillo scope window Similarly adjusting the values manually updates the values displayed in the CONFIGURATION MENU items 3 1 1 Analysis Thresholds As described in Chapter 4 1 the bottom echo of each ping is isolated using two thresholds the bottom detection threshold and the peak threshold You can adjust the thresholds manually by clicking and dragging the horizontal dotted lines in the Oscilloscope window where peak threshold is the top line 2A Peak Threshold Bottom Detection Threshold vr Phe presen E Tin M a Figure 3 1 Bottom Sampling Thresholds in the Oscilloscope Window 3 1 2 Sampling Windows There are several sets of sampling windows represented by vertical lines in the Oscilloscope window When the sampling windows are defined us BS amp E 2005 03 0022 2 2 30 56 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Oscilloscope Window ing Configuration lt Bottom Sampling Windows the user is defining the width of the sample window in units of samples not the placement of the sample window on the horizontal axis The following figure illustrates the
52. VBT 1 10 Visual Bottom Typer User s Guide Janusz Burczynski PhD Catherine Christiaanse M Sc Marek Moszynski PhD BioSonics Inc 4027 Leary Way NW Seattle WA 98107 USA www biosonicsinc com support Qbiosonicsinc com Copyright G 1998 2005 BioSonics All Rights Reserved PRE RELEASE DRAFT C BioSonics Inc 1998 2005 Editors Assad Ebrahim M Sc and Catherine Christiaanse M Sc The authors and BioSonics Inc have taken care in the preparation of this book but make no expressed or impied warranty of any kind and assume no responsibility for errors or omissions No liability is assumed for incidental or consequential damages in connection with or arising out of the use of the information or programs contained herein Typeset by Catherine Christiaanse M Sc using IAIFX Documentation System All rights reserved 11 PRE RELEASE DRAFT Contents End User License Agreement X Installation Instructions XV Acronym Guide xix 1 Getting Started 1 Dil 10 600 3 3 as 4 9 a X xx cx 1 12 Data Collection Guidelines 2 13 Using VBT An Overview 000040 3 1 4 Sample Library aa ara RED EV pd 8 A o AE BG GE e UE A eade ds 9 LO VET 25e 12 0 0 PER EE ERO ESI 14 od Fle Butto 2 a e ab 0 0 2 0 0 E 9 14 W622 Edit BUON 222 exor m x EGENT 6 Ba os 14 1 0 9 Zoom DULLODS be 5 saw 22 0 chew 15 LOA Play
53. XV PRE RELEASE DRAFT xvi INSTALLATION INSTRUCTIONS Installation MM Visual Bottom Typer will be installed into the BioSonics directory of your C drive eABioS onicsWBT A Disk space needed Mb Available disk space 7946 Mb Click Next to continue with this wizard 5 Click Exit to cancel the installation BioSonics 5 select Start to begin installation Installation will take a few seconds Final Confirmation xi ii This program will install Visual Bottom Typer into c BioSonics VBT Click Start ta install Visual Bottom Typer Back to make changes or Exit to cancel the installation BioSonics 6 Select Exit to finish installation and start BioSonics Inc VBT SeaBed Classifier If there were no errors you have successfully installed VBT Seabed Classifier on your hard drive Finish a Visual Bottom Typer has been successfully installed Uninstallation To remove Visual Bottom Typer from your computer use the Uninstaller provided NOTE The User s Guide requires a PDF Viewer If needed you may install Adobe Acrobat Reader from this Installation CD or from www adobe com View User s Guide pdf IM Launch Visual Bottom Typer 7 You can launch VBT using the shortcut on your desktop The VBT program User Guide End User License Agreement and Uninstaller PRE RELEASE DRAFT XV Program can be found in your start
54. after the bottom type as well as the sediment layer In addition the echo itself causes subsequent bottom echoes BioSonics Inc VBT SeaBed Classifier determines the top of bottom by first separating the components of the echo signal sediment echo first bottom echo later divided into two parts and second bottom echo Once the components of the echo signal have been isolated the VBT determines the bottom type according to the user specified signal analysis method VBT includes four independent analysis methods each of which classifies the bottom using different parts of the echo signal Algorithm based classification is a sophisticated subject with a wealth of techniques available to the investigator For this reason the BioSonics Inc You may contact BioSonics Inc directly for further discussion of proper field method ology BS amp E 2005 03 0022 2 2 30 3 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started VBT SeaBed Classifier signal analysis software can seem overwhelming to the beginning user In this chapter you will become familiar with the various windows views and commands needed to perform bottom classification Later chapters will provide hands on tutorials in classifying real data and exporting the results in an import friendly format There are four principal viewing windows in which you may explore viewing your data and the r
55. along bathymetric isolines wherever possible with the transducer central axis adjusted to maintain orthogonality to the bot tom slope e Additional data should be collected in areas of steep slope or that these be surveyed with the transducer central axis adjusted to maintain orthogonality with the bottom slope BioSonics Inc manufactures the DT X 9 with integrated orientation sensor for this purpose BS amp E 2005 03 0022 2 2 30 2 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Using VBT An Overview An in depth discussion of proper field methodology is beyond the scope of this guide We recommend the following data acquisition parameters when working with bottom typing methods B2 B3 and B4 Pulse Duration 4 ms Sampling Frequency 120 kHz We recommend the following data acquisition parameters when working with bottom typing method B1 Pulse Duration 2 ms Sampling Frequency 240 kHz V Note Echo signal parameters depend not only on bottom type but also on the data acquisition parameters Le transducer frequency beam width pulse length etc As such the ground truth data are valid for a particular set of equipment and for particular equipment parameters Ex trapolating to different equipment and or parameters should be accompanied by additional testing 1 3 Using VBT An Overview The echo signal from an underwater acoustic pulse contains information
56. are and your registration codes and license keys for back up purposes only provided that this back up copy contains the same proprietary notices as appear in this Software and provided that it is adequately protected by you In particular at no time should your registration codes or license keys be stored with the Software in a publicly accessible or networked environment You may not modify adapt translate reverse engineer decompile disassem ble or create derivative works based on the Software You may not modify adapt translate or create derivative works based on the written materials without the prior written consent of BioSonics Inc This Software is licensed to only you the Licensee and may not be trans ferred to anyone else without the prior written consent of BioSonics Inc In no event may you transfer assign rent lease sell or otherwise dispose of the Software on a temporary or permanent basis except as expressly provided herein However PROVIDED that you DO NOT include your personal registration codes or license keys you may distribute copies of the auto installation media containing this Software or the accompanying written materials to be used by others for their trial Warranties and Warranty Disclaimer THESOFTWARE AND ACCOMPANYING WRITTEN MATERIALS ARE PROVIDED AS IS WITHOUT WARRANTY EXPRESS OR IMPLIED OF ANY KIND AND BIOSONICS INC SPECIFICALLY DISCLAIMS PRE RELEASE DRAFT THE WARRANTIES OF FI
57. asses i e mud and mud with sand Figs 7a and 7c The separation between these bottom classes is rather BS amp E 2005 03 0022 2 2 30 D BioSonics Discussion good However this was to small data set only two bottom types in order to draw conclusion on influence of pulse du ration on bottom classification results for 420 kHz It will be necessary to ac quire more data on this topic The worst separation occurs in all of the cases for the rocky bottom It is caused by the highest value of the pa rameters scattering in the case of this type of the bottom This scattering is the result of the big dynamics of the sig nal changes even in the case of slightly different incidence angles related to the bottom unevenness Comparing classification results for the pulses with maximum amplitude with the averaged ones for all of the pulses it can be noticed that selection of the pulses with maximum amplitude improves differentiation in case of the method with division of the first echo signal It is vis ible both in all frequencies 38 kHz 120 kHz 420 kHz In the case of the second echo method for 38 kHz and 20 kHz it can be seen that differentiation is improved for the averaged results though averaging caused visible in both methods decrease of rough ness and hardness measures as well as the dynamics suppress This change is caused by the contribution in the aver aged results of pulses scattered at oblique
58. atitude of the ping transmission Longitude longitude of the ping transmission Depth depth to the bottom Type numerical code corresponding to the type of bottom no bottom coded corresponds red parameter box first file opened corresponds to the green parameter box second file opened N L O corresponds to the blue parameter box third file opened BS amp E 2005 03 0022 2 2 30 5 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started EO energy of sediments echo El energy of second part of 1st bottom echo E2 energy 2nd bottom echo El energy of first part of 1st bottom echo Sediment thickness of the sediment layer FD Fractal dimension See Section 3 4 for more information on working with the output report window A fourth window map window can be opened manually by the user by selecting the Show Map Window command from the Vrgw MENU or EDIT MAP MENU Map images can be imported into VBT using the EDIT MAP Menu located in the menubar See Section 3 5 for more information on working with the map window Figure 1 4 An example of the map window from sample library file TR 4 A fifth set of windows referred to as the method windows allows the user to customize the settings for the bottom typing method in use There is one method window for each bottom typing method B1 B2 B3 B4 S See Section 3 2 for more i
59. averaged ping values for every report from the current data file The red dots signify average values Multifile View allows you to view the average ping values for every report for all open files The dots are Color coded in the order the files were opened red lst open green 2nd open blue 3rd open FCM Algorithm Refer to the following section 60 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Method Windows 3 2 8 Fuzzy C Means FCM Clustering The FCM Algorithm allows you to automatically create the feature space boxes contained in the B2 B3 or B4 method windows The algorithm should be used when generating your ground truth reference library See Section 2 1 1 for a tutorial on working with the FCM algorithm Fuzzy C Means Clustering BioSonics Inc x Humber of clusters Database option Cancel C Display cluster centers only Append to current database Horizontal axis multiplier Vertical axis multiplier B References Rectangle size Bezdek J C 1881 Pattern Recognition Algorithms with Fuzzy Objective Functions Plenum Press New York Figure 3 4 FCM Algorithm Options Number of Clusters specifies the number of clusters for the algo rithm to locate The number of clusters should always be equal to the number of data files used to generate the ground truth data Display Clus
60. c All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter2 Tutorials Fractal Dimension B4 Method BioSon ics T ic lipdate Close Lin Log lv Rescale gt bs info gt gt ED import Export Table 2 2 Recommended Bottom Sampling Window Widths lll Bottom Sampling Windows BioSonics Inc E1 First Bottom First Part samp El First Bottom Second Part 48 samp Cancel E2 Second Bottom Window samp Sediment Window zamp Data Acquisition Pulse Duration 6 samp Step 6 Go to Configuration gt Output Report Filters Select stop analysis after one report and leave all other fields with their default values as shown below Configure Analysis Output Report BioSonics Inc E Energy filter Z pou 2 Pings per Report Cancel M Report the average of all qualifying pings V Stop analysis after one report Play Step 7 Begin analysis of the data file by selecting the play button in the toolbar After each report VBT pauses analysis and all pings will be displayed in B4 method window o Red Dot Averaged value for all pings that pass through the filter o Blue Dot ping with more than 60 of energy BS amp E 2005 03 0022 2 2 30 44 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Ground truth Library o Green Dot ping with less than 60 of ener
61. ch vertical line corresponds to a lost bottom In order to reduce the number of lost bottoms we recommend first lowering the Alarm Limit found in Configuration lt Bottom Tracking Parameters and re analyzing the data by selecting View lt go to first ping and Analyze lt Play If the number of lost bottom persists go to the oscilloscope window and en sure that the bottom sampling windows defined in Configuration lt Bottom Sampling Windows are properly positioned Please refer to Section 4 1 for more information on working with the Bottom Tracking Algorithm Untyped Pings If a large number of the pings in your data file are not assigned a bottom type during analysis it is most likely due to small feature space boxes in the method window When you are creating a ground truth data set be sure that your feature space boxes fill as much as the feature space as possible All pings that fall outside of the boxes will be un typed BS amp E 2005 03 0022 2 2 30 53 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter2 Tutorials 2 5 Exporting the Analysis Results Please refer to Section 1 7 4 for information on exporting the results of the analysis 2 6 Analyzing Batch Files 2 GIS Ready Analysis BS amp E 2005 03 0022 2 2 30 54 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Chapter 3 Manipulating VBT Windows
62. customized to the region you are studying The feature space is the white region in the method window and each feature space box represents a different bottom type In VBT you create additional classification sets by clicking on the New BS amp E 2005 03 0022 2 2 30 58 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Method Windows Feature Space Feature TUE ED 1 10001 Expar 1120 kHz 0 4 ms SN rage Bolwin 16 48 PL 16 M DT20 DT4 red Bot Type 1 SAND120 074 green T a ROCK120 DT4 blue T3 button in the method window you wish to use You can annotate the classifi cation set by clicking on Info and adding pertinent details such as location sampling frequency and pulse duration 120 kHz Dd ma 5H ESHSH Ho won 16 4B PL ih DOT 4 red Hot Type 1 SAMDT71 T4 giran T2 M FHICK 121 D T 4 The feature space is numbered in the or der in which they were created The pa rameter set in use is identified by the ID number highlighted in the accompanying figure If you select a parameter set prior to analysis VBT will classify the bottom using the parameters you selected Oth erwise VBT will classify the bottom us ing the last parameter set you used You sum Bones can move between feature space sets that 4 Tf mran 2L Hl t Ta Ban EE you created by clicking on the right and left arrows buttons
63. d 50 bthickness 5 sthreshold 69 sthickness 1 alarm 8 bwindow 66 domain 0 TVG 0 depth 1 lin 0 blength 48 blength2 100 subbottom 50 bottomUffset 16 initialoffset 0 grid 0 ticks 1 threshold 1 fillbottom 1 bottoms 1 CalcBottom 1 B2rescale 1 B3rescale 1 StopIt 0 SimradScope 0 DutputDateFormat 1 LatLonFormat 0 oeparator 0 colorbar 70 3 30 Results window size Rect 002 0264 0770 0401 icon 0 max 0 Scope window size Rect 003 0008 0342 0262 icon 0 max 0 Colors Templates 255 65280 16711680 16711935 8388 36 8388736 8388736 BS amp E 2005 03 0022 2 2 30 108 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 8388736 8388736 8388736 16711935 8388736 8388736 8388736 8388736 8388736 Echo 0 0 O Grid 0 O 255 Chart 255 255 255 Level 0 128 0 Bottom 128 O 128 Bottom2 128 0 0 Current 0 64 4210816 255 8421631 16384 65280 8454016 4210688 8404992 16776960 16777088 8553090 12632256 16777215 16777215 Current 0 4210816 255 8388863 16711935 4227327 65535 32768 65280 65408 8388608 16711680 16776960 16777215 16777215 16777215 Palettized 0 Rect5Library01 No 3 Recti 41 75 97 122 Info00 120 kHz 0 4 ms SN 49408 BotWin 16 48 PL 16 Rect2 90 33 138 79 Info01 MUD120 DT4 red Bottom Type 1 Info02 SAND120 DT4 green T 2 ROCK120 DT4 blue T3 Rect3 132 16 171 76 Rect6Library01 No 3 Rect1 8 91 99 151 Rect2 72 59 142 93 Rect3 113 28 176 6
64. d gt ROCKT20 DT4 We c Ur Horizonatal axis multiplies 2 Vertical axis multiplier 25 Results filename fem out Figure 4 11 Fuzzy C Means Clustering technique for different bottom cat egories Fractal Dimensions method B4 for data files from VBT Catalogue mud120 dt4 sand120 dt4 and rock 120 dt4 BS amp E 2005 03 0022 2 2 30 82 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Appendices Editor s Note The scientific papers included in Appendices A B amp C are translated from Polish into English are were only edited for spelling and basic grammar The original integrity and writing style were maintained C JC PRE RELEASE DRAFT VBT User s Guide 2 Chapter4 How VBT Works BS amp E 2005 03 0022 2 2 30 84 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Appendix A Method B1 Publication Method Of The Sea Bottom Classification Using Bottom Echo Cumulative Energy Curves vspace lin Dariusz Bakiera Technical University of Gdansk A 1 Introduction Pouliquen and Lourton proposed a method of sea bed classification using cumula tive curve of echo signal envelope They assume the shape of the cumulative en ergy curve should have a distinct shape and it should be different for different bottom classes The parameter used to describe the shape of the echo envelope is a cumulative sum of energy of an echo
65. d the algorithm searches for the next bottom within the window in both directions Lost Bottom When a bottom echo is not found within the bottom tracking window it is referred to as a lost bottom BS amp E 2005 03 0022 2 2 30 26 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Menubar Commands Configuration lt Select Bottom Typing Method To change the bottom typing method used to classify your data set choose Select Bottom Typing Method from the CONFIGURATION Menu This is the same as selecting the bottom typing method in the toolbar Method B4 is the default and recommended bottom typing algorithm for use with VBT and will be implemented automatically when you do not select an alternative bottom typing method lll Bottom Typing Algorithms BioSonics Inc Advanced Digital ydroz 6 pusiese xj Select the principal bottom typing algorithm to be used B1 First Echo Normalization Lourton Pouliquen Cancel The method of cumulative curves B2 First 5econd Bottom Echo Ratio E1 E2 Orlowski Chivers The RoxAnn method C B3 First Echo Division E1 E1 Stepnowski Bakiera Moszynski The BioSonics method 1 BA Fractal Dimension Lubniewski Stepnowski Moszynski The BioSonicz method 2 Configuration gt Configure Bottom Typing Algorithms You may open up the 1 Configuration method window for any Data File Transd
66. e bimodal distri bution of box dimension values for echo envelopes for rocky bottom for the data from transects Fig 5 This form of distribution is due to a recording made in transects of alternating series of pulses of a twofold shape The examples of os cillograms of two classes of echo pulses from the transects are presented in Fig 6 The existence of these two echo pulse types may be explained by alternative reception of echo signals reflected at nor mal incidence specular reflection and echo pulses received at oblique incidence The first type pulses have higher ampli BS amp E 2005 03 0022 2 2 30 D BioSonics ChapterC Method 4 Percentage of pings 3m 3 Zo zr T 0 5 Box dimere ion Box dimers ion Har and Softeand Figure C 3 The histogram of the box dimension values eval uated for echo pulses from sta tionary data for the echosounder frequency 120 kHz a for all echo pulses b for selected 10 of echo pulses of the highest am plitude level tude and less fluctuating envelope shape against the higher amplitude and more fluctuating envelope shape of the second type pulses see Fig 6 In the case of the 38 kHz echosounder frequency Fig 4 where the only data available was data from the anchored ship there is a clear division between box dimension values obtained for vari ous types of bottom However the mutual relations between
67. e en ergy curves obtained for 120 kHz are less distinguishable from each other for dif ferent bottom type than curves obtained in each set of bottom class In this case bottom classification would be not pos sible with this method Better results were obtained for 38 kHz However for 420 kHz the standard curves have dis tinct different shape for different bottom classes For 420 kHz curves for sand and sand with mud have a similar shape while remaining tow classes of the bot tom i e rock and mud produce curves of distinct different shape A 4 Reference E Pouliquen and X Lourton Sea bed classification using echo sounder signals Proceedings of European Conference on Underwater Acoustics Luxembourg Sept 1992 535 538 LIST OF FILES a 38 kHz rock38w1 dt4 sand38w1 dt4 ssan38w1 dt4 smud38w1 dt4 b 120 kHz BS amp E 2005 03 0022 2 2 30 D BioSonics ChapterA Method 1 hardsand dt4 rock dt4 softmud dt4 softsand dt4 c 420 kHz mud420a dt4 mud420b dt4 rock420a dy4 samd420a dt4 samd420b dt4 sand420a dt4 d 420 kHz PEPIN files from Bruce Sabol 11 files pepin_1 dt4 do pepin_11 dt4 88 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Methods Method Of The Sea Bottom Classification With A Division Of The First Echo Signal And Using First Second Bottom Signal Ratio Dariusz Bakiera Andrzej Stepnowski Technica
68. e threshold in order for the dip or jump to be classified as an end or peak Without ike Abore Bokom Slanking core this dip would Without tre Peak Width this jump would be incorrecthy be incorrecthy Peak Width a 4 identified as the end identified as the ofthe bottom echo Er peak of the bottom echo Peak Threshold T gt F E Above Bottom Blanking Zone E t Bottom Detection Threshold S amples Axis Fish Echoes Bottom Echo Figure 4 2 Bottom Tracking Sample Widths The sample widths prevent the algorithm from incorrectly classifying small jumps and dips as the bottom echo peak or bottom echo endpoint respectively Once a bottom echo has been identified VBT finds the next bottom echo by searching for samples that rise above the peak threshold within a user defined bottom tracking window If the bottom peak is not found within the window then the bottom is considered lost The previously found bottom is assumed to be this ping s bottom If the number of consecutively lost bottoms exceeds a fixed number alarm limit then the bottom tracker is re initialized and begins searching from the absolute bottom of the ping rather than the bottom of a window The bottom tracking parameters can be defined by the user in Configura tion lt Bottom Tracking Parameters lll Bottom Tracking Parameters BioSonics Inc Bottom Detection Peak Threshold
69. ea for the surface reverberation and disturbance for volume reverberation Examples of model standard curves ob tained by Pouliquen and Lourton are pre sented on the Fig 2 The purpose of this report is to describe an attempt to create standards which may be used in bottom classification from known experimental data e g recorded bottom echoes for verified bottom sam ples A 2 Results Echo signals were acquired with follow ing frequencies 38 kHz 120 kHz and LEASE DRAFT VBT User s Guide 2 39 1732 Figure A 2 Set of normalized and integrated theoretical ampli tude envelopes E t H 50 m directivity of the transducer 16 frequency 38 kHz by Pouliquen and Lourton 420 kHz For the 38 kHz and 120 kHz frequencies the bottom echo signals were recorded for the following bottom types rock sand soft sand and soft mud For 420 kHz frequency the bottom echo sig nals were acquired for rock sand mud and mud with gravel In order to assess the influence of oblique reflection from the bottom each series of data was analyzed in two ways First by taking into account only perpendic ular echoes and and second averaging all echo signals for given class of the bottom Echoes were filtered by select ing local maxim of amplitude Those echoes containing local maxima were as sumed to be formed by a perpendicular reflection from the bottom For 420 kHz data two data sets were available
70. ed by Pouliquen and Lour ton 1 This method is quickly becoming obsolete and is included in the man ual for historical purposes This signal processing technique characterizes the signal by sampling the 1st echo signal from the bottom with a wide window several pulse lengths and calculating the cumulative energy curve of the echo The cumulative energy curve is built for each echo signal cumulative time integral of squared echo envelope and averaged over a user specified number of pings Different types of the bottom e g mud sand rock will produce curves of different shapes A hard bottom will produce a sharp bot BS amp E 2005 03 0022 2 2 30 5 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter4 How VBT Works tom echo with high amplitude while a soft bottom will produce an elongated echo with lower amplitude The cumulative curve of an echo signal acquired during survey is then com pared with verified standard curves The bottom type of an unknown signal is determined by taking the least squared error between the unknown curve and previously classified curves Figure la displays examples of echoes from hard and soft bottom Figure 1b displays the integral of the squared echo hard br4tom A f hard boltom i amalitude tima capis 4 meras Figure 4 3 Different shapes for hard and soft bottom a Echo Signal Am plitude b
71. eflection from lossy bottom 1 6 analysis of a set of values of acoustic and statistical parameters of the echo envelope of reverberation sig nals using algorithms of neural net works or cluster analysis 4 This paper describes rather simple method of sea bed classification which can be implemented using single beam echosounder and a personal computer B 2 Method of sea bottom classification with the first echo division Classification of sea bottom can be done by estimating two parameters encoded in the echo signal envelope hardness and roughness The parameters can be extracted by taking of the integral of sea bed echo envelope in respective time intervals Information on hardness and roughness of the bottom is embedded in the part of echo signal related to back scattering on the boundary water bottom In order to perform bottom classifica tion it is necessary to split the two com ponents of an echo signal i e one caused 89 PRE RE LEASE DRAFT VBT User s Guide by a surface reverberation hardness sig nature and the other caused by a vol ume reverberation roughness signature The volume reverberation can be treated as a disturbance which should be min imized The attempt to split two sig nal components from the first bottom echo signal were done based on numeri cal simulation results of echo signal en velope back scattered from the bottom 1 Models of sea bottom echo envelopes comp
72. eport stops the analysis after 1 report If report the average of all qualifying pings is selected the analysis will stop after 1 ping report and the output report window will show 1 report containing the average of all Pings per Report pings that exceed the energy filter If the averaging option is not selected all pings that exceed the Energy filter are reported for each grouping of Pings per Report In either scenario the next report is generated by pressing the play button This option is extremely useful when customizing feature space using the method windows BS amp E 2005 03 0022 2 2 30 65 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter3 VBT Windows vm x The format for the date longitude amp latitude and output file are speci fied in the accompanying window by selecting Configuration lt Output Report Format Options Specifi Date Format C mm dd yy dd mm yy Latitude Longitude Format i NMEA Numerical Output Fields Separation Symbol cations made here will affect both the Comma Tab output report window and any ex ported report file Recall changes If the changes are made after play has been used the effects will take place when the file is reanalyzed You can specify the data fields to include in the output report window and any related export files by going to Configurat
73. er s Guide 2 Map window itude and longitude The highest latitude is defined in the top box The west most longitude is defined in the left box You can also add coordinates by double clicking on the map window and entering them into the pop up menu However it is recommended that you use the Calibrate Spatial Range Command to ensure accuracy If a map is im ported from EchoBase then the spatial range is entered automatically If a map window has been opened prior to analysis the bottom types will be labeled in the appropriate coordinates directly onto the map file during analysis The bottom types are defined by the parameter sets defined in the Method window and labeled with colored dots consistent with the color code in the Method window Unclassified bottom types are identified by white colored dots TIP If the map window is opened after analysis has begun go to the first ping and reanalyze the data VBT allows you to use any bitmap or windows metafile file as a map file It is up to the user to acquire an appropriate image BS amp E 2005 03 0022 2 2 30 67 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter3 VBT Windows BS amp E 2005 03 0022 2 2 30 68 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Chapter 4 How VBT Works This chapter provides a brief description of the theory behind the algorithm in VBT T
74. er s Guide For this set of 420 kHz data both first and second bottom echoes were recorded and therefore analysis was done using first bottom echo division method and also first second bottom echo ratio Fig 7 B 4 Discussion As indicated in the figures it was not possible to obtain a complete separation of areas occupied by different classes of the bottom in any of the cases For both methods first echo division and first second bottom echo ratio the clas sification results are better for the fre quency 120 kHz This effect is probably due to less contribution from volume re verberation in this kind of signal at 120 kHz than at 38 kHz In figures 5d and 6d we can see good separation of follow ing types of the bottom soft sand soft mud and sand For low frequency 38 kHz and also for high frequency 420 kHz we can ob serve improvement in separating the most class of the bottom with the pulses of longer duration 7 0 4 ms However harder types of the bottom were more difficult to separate In the case of 420 kHz data set the reason may be that the bottom classes were not the same as for 38 kHz and 120 kHz soft mud soft sand while mud and mud with sand were used at 420 kHz Scattering dia grams demonstrate that the bottom mud with sand has similar properties to rock and sand bottom rather than to soft sand Results for 420 kHz with pulse dura tion 7 0 4 ms correspond only to two bottom cl
75. es to the database you just created y Note When creating a ground truth library record 1 the order in which the feature spaces were created and 2 the data acquisition parameters in the Info section for future reference Step 12 b Set the Number of Clusters to 3 The number of clusters should always be equal to the number of data files in use Step 12 c Set the Rectangle Size multipliers to 3 The multi plier expands the box size in multiples of the variance of each cluster set T he default value is one BS amp E 2005 03 0022 2 2 30 41 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter2 Tutorials Step 13 Close the FCM Window You should see a circle and a rectangle around each data clusters The circle represents the space created by the FCM algorithm and cannot be altered The rectangle is the fea ture space generated from each circle The rectangles can be resized manually using your mouse C Lin Log V Rescale ID Type ID amp Step 14 Hit Update in the method window Update stores the database The database will not be stored if without selecting update Step 15 Close the method window and follow the steps outlined in Tutorial 2 1 3 using the database you just created BS amp E 2005 03 0022 2 2 30 42 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT
76. eserved PRE RELEASE DRAFT VBT User s Guide A Analyzing Your Data File ii Select the the corresponding method from Select Bottom Typing Method in the CONFIGURATION MENU iii Select the the corresponding method from Configure Bottom Typing Algorithms in the CONFIGURATION MENU Step 4 Select the ground truth library you wish to use by cycling through the ID buttons y Note When creating a ground truth library record 1 the order in which the feature spaces were created and 2 the data acquisition parameters in the Info section for future reference DA LHo 8d ma SH ABEJA Barwin 16 48 PL 16 ODT DT 4 red et Teen 1 AMDTZ DTE een 12 CK A TA Els T 1 otep 5 You are now ready to begin analyzing your data file 2 3 Analyzing Your Data File 2 3 1 Single Transducer Data Files Once you have prepared the data file properly as described in Section 2 2 you may analyze the data file in one of three ways Play e Press the play button located in the toolbar e Select Play from the ANALYZE MENU e Press F5 2 3 2 Analyzing Multiplex Data Files BS amp E 2005 03 0022 2 2 30 51 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter2 Tutorials Only one channel can be processed at a time VBT will open the lowest channel number by default Channel 1 in this example You can select the channel you wish to
77. esults of VBT s bottom classification analysis i oscilloscope windows ii echogram windows iii output windows iv map windows When a data file is first opened in VBT it is displayed in an oscilloscope window Figure 1 3 Oscilloscope windows show the echo envelope for all pings included in the data file one ping at a time Depth range is shown along the horizontal axis with signal intensity in decibels along the vertical axis The headerbar indicates the number of the ping being displayed the total number of pings in the data file and which transducer channel is active The data file name is displayed in the title bar See Section 3 1 for more information on working with the oscilloscope window STAT 4 Chi 1075 Figure 1 1 An example of the oscilloscope window from sample library file TR7_B dt The title bar displays the file name The heading below Ch 1 Ping 61 978 indicates that the 615t ping out of 978 pings from Channel 1 is displayed When the data is being processed two more windows open the echogram window and the output report window The echogram window displays all of the pings in the data file Each vertical line corresponds to a single ping and the color of each line represents the echo intensity Depth is shown on the left hand vertical axis the intensity colorbar is displayed on the right hand vertical axis the top horizontal axis corresponds to the ping number and the bottom h
78. fferent scales 0 005 of 018 02 025 03 0 04 on graphs on the Fig 8 El measure ofrougness Regions with data points for different Figure B 9 Classification groups with partial overlap make bot results for s of pe tom classification more difficult How ent grain size frequency heth h kHz pulses with maximum am ever we can not say whether the rea E Oe a son is the deficiency of the method or of first echo b second echo whether just samples of bottom classes method o coarse grain fine are overlapping IT is likely that bottom grain on similar grain size produces acoustic samples with hardness amp roughness sig nature close to each other e g samples 11 One can expect that those signals 1 10 or 5 6 7 Fig 9 presents PEPIN correspond to the hardest and most un A even bottom surface Experiments showed that the method with the division of the first echo signal gives comparable results Similar as to data analyzed for 38 kHz With those obtained from the second echo and 120 kHz also for 420 kHz frequency method In the case of a normal inci data signals of larger amplitude have lager dence the results are even better Its ad variability e g bottom samples 9 and vantage is that it does not need the sec ond echo signal which can be obtained data pints divided into two classes of the bottom fine grain and coarse grain BS amp E 2005 03 0022 2 2 30 98 Copyright C 1998 2005 BioSonics I
79. fied file This function is only available for the Echogram window the output report window or the map window The window type and its corresponding file format are listed in the Table 1 7 4 y Note if the window has been zoomed into only the zoomed in portion of the window will be copied Zoom out to copy the entire window BS amp E 2005 03 0022 2 2 30 32 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Menubar Commands 1 7 5 Edit Map Menu Edit Map gt Show Map Window The Show Map Window command allows you to associate a map image with the open data file A map image file is associated with a data file when the map picture file is opened while the data file is open by using either Import Map Image from File or Paste Map Image from Clipboard from the EDIT MAP MENU Once a picture file is associated with the data file it can always be opened using the Show Map Window command Edit Map gt Import Map Image from File In order to import a map image from a file the Show Map Window com mand must first be selected VBT supports map images of the following formats bitmap bmp Windows Metafile wmp and Graphic Inter change Format gif y Note VBT is indiscriminate with respect to map files If you import any supported graphics file VBT will assume the file is a map of appropriate coordinates and plot the results accordingly Edit Map gt
80. g pulse propaga tion for an ideal beam pattern For long pulses and for narrow beam transducers Phase II will be shorter un til it disappears completely see Fig 1 b In the case of an ideal beam pattern after the end of the release phase re verberations due to bottom volume in homogeneities are the only source of the echo This phenomena is prevalent in 90 Copyright 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide DD First Echo Division the case of soft types of the bottom for which the main part of the echo energy penetrates the bottom Fig 3 illus trates the model echo envelope gener ated by a soft bottom Figure B 3 The results of echo envelope simulation for soft mud in case of an ideal cone beam pattern of 26 width bottom surface reverberation bottom volume reverberation sum In the case of a real beam pattern with a side lobes the part of the signal that reaches the transducer after the ending of the release phase is due to bottom sur face scattering It is however received mainly by the side lobes of the beam pattern and the reception is done at a wide angle so its level is negligible ac cording to Lamberst s law The real beam pattern influence appears also in spreading of decay and release phases limits see Fig 4 Figure B 4 The results of echo envelope simulations for bottom types fine sand
81. gram Color Bar To change the scale or range lll Echogram Amplitude Colar Bar Options BioSe nic n Inc xX of the echogram color bar se lect Echogram color bar in the amum value p CONFIGURATION MENU when the step a Cancel Oscilloscope window is active The Ge E pop up window allows you to change the minimum maximum and step value in decibel units displayed on the colorbar on the right hand side of the Echogram window Each step is assigned a new color from the color table defined in Configuration gt Colors gt Echogram Configuration gt Colors Selecting Colors in the Con Configuration FIGURATION MENU brings up a Data File Transducer Properties Ctrl Enter submenu The first five options 9sciloscopes Echagram Color Bar Scope Ping First Bot EEE lt lt tom Windows Second Bottom o Bottom Sampling Windows ai 4 Em M um Window and Ihresholds al Bottom Tracking Parameters s dubius low the user to change Colors 5 Bottom Typing Method Configure Bottom Typing Algorithms b Echogram associated with the Oscilloscope ur window Mark Bottom Bottom Types This function does not work properly for this release of VBT BS amp E 2005 03 0022 2 2 30 23 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started Scope allows you to select the background color f
82. gy See Section 3 4 for an explanation of the energy filter Step 8 In the method window draw a box around the blue dots pings that exceed the energy filter using the mouse An example of the first report display is shown below lll Fractal Dimension B4 Method BioSonics Ine Info gt gt ED import Export Step 9 Adjust the feature space red box to tightly fit around all of the blue dots Click Update This saves the settings to the catalog so it can be used to classify unknown data sets in the future Fractal Dimension B4 Method BioSonics ine Step 10 Play the next ping data set and again adjust the feature space to encompass all of the blue dots Be careful to adjust only the size of the box and not to move the box or you may lose information from the last set of pings Click Update again Step 11 Repeat the cycle analyze adjust the box and update the settings until the entire data file as been processed Step 12 Right click on the titlebar and select Report View This will show a cluster of red dots signifying the average value of all reports contained within the data file Ensure the feature space encompasses all dots Click Update BS amp E 2005 03 0022 2 2 30 45 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter2 Tutorials Step 13 Step 14 Tip Be sure to select stop analysis after one repo
83. he algorithm for VBT can be divided into two main parts 1 the bottom tracking algorithm and 2 the bottom typing algorithms In order to perform bottom typing VBT first identifies the bottom echo portion of the data file using the bottom tracking algorithm described in Section 4 1 Once the bottom echo portion has been isolated VBT uses the Bottom Typing method selected by the user in order to determine the bottom type The bottom typing methods included in VBT are described in Section 4 2 A complete discussion of the algorithms included in VBT is beyond the scope of this User s Guide A sample of the published papers regarding bottom typing algorithms are provided in the appendices for your reference 69 PRE RELEASE DRAFT VBT User s Guide 2 Chapter4 How VBT Works 4 1 Bottom Tracking The purpose of VBT is to classify the type of bottom of the seabed by analyzing the signal characteristics of the bottom echo In order to classify the bottom the portion of the bottom echo portion of the ping must be isolated VBT uses an algorithm called the Bottom Tracking Algorithm or Bottom Tracker to isolated the bottom echo BioSonics Inc bottom tracker is an end up algorithm in that it begins searching for the bottom echo portion of a ping from the last sample toward the first sample T he bottom tracker tracks the bottom echo by isolating the region s where the data exceeds a peak threshold for
84. he feature space lutorial 2 1 1 demonstrates using the FCM algorithm and Tutorial 2 1 2 demonstrates manually generating the feature space The results of both of the tutorials can be used in Tutorial 2 1 3 to classify your unknown data set The following tutorials use Method B4 but can also be applied to Methods B2 amp B3 y Note IT IS IMPORTANT TO FOLLOW THESE IN STRUCTIONS PRECISELY PARTICULARLY THE OR DER OF TASKS 2 1 1 Creating a Ground truth library using Fuzzy C Means FCM Clustering VBT allows you to automatically generate feature space boxes for bottom typing using the FCM Algorithm Please see Section 3 2 3 for more informa tion on working with the FCM Algorithm in VBT Step 1 Open data file mud120 dt4 Step 2 Activate method B4 by clicking on B4 in the toolbar or selecting BS amp E 2005 03 0022 2 2 30 38 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Ground truth Library Configure Bottom Typing Algorithms B4 from the CONFIGURATION MENU Step 3 In the method window select the New button to generate a new database This will create a ground truth data set p ies S00 o i bl 120 kHz 0 4 ms 5H 49408 Botwin 16 48 PL 16 M0D120 074 red Bot Type 1 SANDT20 DT4 green T2 H CK120 DT4 blue T3 Close C Lin Log gt Rescale Type ID Fractal Dimension B4 Method BioSonh Step 4
85. ht C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide Your Data File for Analysis The methods described in the above tutorial should be used to establish a ground truth for each of your own data sets Recall it is important for un classified data sets to be collected using the same data acquisition parameters that were used to collect the ground truth data 2 2 Preparing Your Data File for Analysis Once you have generated a ground truth you are ready to begin classifying your unknown data sets The first step in classifying your unknown data set is preparing the data file for analysis 2 2 1 Oscilloscope Window The bottom tracking algorithm is the first step in the classification process in VBT The bottom tracking algorithm isolates the bottom echo portion of the each ping See Section 4 1 Before you begin analysis of your data you need to set up the bottom tracking parameters in the oscilloscope window The threshold values and sampling windows may be adjusted either manu ally or numerically The values can be defined numerically by goingto Configuration lt Bottom Tip Tracking Parameters for threshold values and Configuration lt Bottom Numerical Sampling Windows forthesampling windows Adjusting the values numeri adjustments cally adjusts the visual display in the Oscilloscope window Similarly adjust affect the width ing the values manually updates the values displayed i
86. incidence from the bottom caused by tilt of the boat and transducer These pulses are characterized with a smaller amplitude than those from the normal incidence Estimation of bottom class for PEPIN data set the one with percentage con tents of particles of different size is rather 97 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 ChapterB Methods B2 amp B3 difficult since there is no available rela tionship between bottom class and the percentage particles contents Although 0 25 we have no explicit description of bot tom samples there are some regularities in relationship between hardness amp rough ness signatures and particles contents E ho El measure ofrouhgness The bottom particles size was divided into three groups coarse grain more 0 05 P NE than 50 wm medium ones with from 2 um to 50 um and fine grain less than 2 0 O05 01 015 02 025 03 035 4 E1 measure af roughness pm The content of fine particles less than 2 um is approximately the same in all samples Than bottom classes can be divided into two classes of percentage content of fine particles coarse grain samples 1 5 6 7 8 10 and fine grain samples 2 3 4 9 11 The bottom classification analysis hard ness amp roughness signature demonstrates that experimental data PEPIN can be grouped into two clusters which was vi f sualized by choosing two di
87. indows You have successfully created a ground truth library Proceed to the next section to learn how to analyze a data file using your ground truth library BS amp E 2005 03 0022 2 2 30 PRE RELEASE DRAFT 47 Copyright C 1998 2005 BioSonics Inc All Rights Reserved Tip Be sure to press the Update button throughout analysis VBT User s Guide 2 Chapter2 Tutorials 2 1 8 Classifying data using your ground truth library Step 1 Open file T R7 B dt4 Step 2 Go to Show Map Window in the EDIT MAP section of the menubar Step 3 If a map file does not open goto Import Map from File andopen C BIOSONICS VBT MAPS MAPA bmp Step 4 Goto Configuration lt Output Report Filters and deselect Stop analysis after 1 ping Step 5 Analyze the data file by pressing O The algorithm will classify each report of pings according to where the averaged ping falls in feature space shown in the method window during analysis O In the map window you will see an array of dots T he colors of dots correspond to the colors in the multifile view setting of the method window In this case the red dots represent to mud the green dots represent to sand the blue dots represent to rock and the white dots represent to unclassified pings The final output should look as follows TR B DT4 2 AT 43 N O In the echogram window you should see the bottom type for each ping color coded along the bottom BS amp E 2005 03 0022 2 2 30 48 Copyrig
88. ing by 60 For example 122 38 2200W converts to degrees minutes seconds by multiplying 22 by 60 which results in 122 38 13 2W Decimal minutes are converted to decimal degrees by dividing the minutes by 60 For example 122 38 2200W converts to degrees decimal minutes by dividing 38 22 by 60 which results in 122 637 1 7 6 Analyze Menu You may begin analysis of the data by 1 selecting Play in the ANALYZE MENU 2 hitting the play button in the toolbar or 3 pressing F5 on your keyboard If Stop analysis after 1 report has been selected via Configuration gt Output Report Properties you may continue analysis by pressing play until you reach the end of the data file If you make changes via the CONFIGURATION MENU after analyzing the data you can re analyze the data by selecting the go to first ping button in the toolbar and hitting play again new Echogram and Output Report windows will open Pressing in stop the toolbar F10 on your keyboard or selecting Stop from the ANALYZE MENU will stop analysis of the data BS amp E 2005 03 0022 2 2 30 34 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Menubar Commands 1 7 7 Window Menu Duplicate Windows creates a copy of all open windows and allows you to reanalyze the data Cascade cascades all open windows Tile tiles all open windows 1 7 8 Help Menu The help command allows you
89. ing step tends to 0 The Hausdorff dimension 3 4 may be the solution of this problem as it can be used as a measure of many very general sets including fractals The Hausdorff dimension of a subset X of Euclidean space is defined as a limit log N aslo 1087 where N r denotes the smallest number of open balls of radius r needed to cover X D C 1 The open ball B p r x dist x p gt r where dist x p is the distance BS amp E 2005 03 0022 2 2 30 D BioSonics ChapterC Method B4 between points x and p For example the Koch snowflake has the Hausdorff dimension equal to log 4 log 3 1 262 qu am x E ud Figure C 1 The Koch snowflake constructed as a limit of a sequence of simple iterative steps Starting with the equilateral tri angle each consecutive stage is constructed by replacing line segments with copies of the figure In the random version in each stage the replacement by A mE may be made with probability of 0 5 for each case It is clear that the dimension as defined above is the measure of the complex ity of a given figure and we could ap ply it to measure the complexity of the shape of a digitized echo pulse from the sea bottom It should be the indicator of the complexity and also the type of sea bottom The shape of a digitized echo pulse as it consists of a finite set of straight sections is not really fractal However
90. ing the average of all Pings per Report pings that exceed the energy filter If the averaging option is not selected all pings that exceed the Energy filter are reported for each grouping of Pings per Report In either scenario the next report is generated by pressing the play button This option is extremely useful when customizing feature space using the method windows BS amp E 2005 03 0022 2 2 30 28 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Menubar Commands Pings per Report N number of pings to be included in one report in the output report window lf averaging is selected the report will show the average of each set of N pings that surpass the energy filter If averaging is not selected all pings that exceed the ENERGY FILTER will be reported independent of the value of N Energy Filter The minimum energy a ping must have to be reported calculated as a percentage of the maximum energy present in the data file Configuration Output Report Export Fields You can specify the fields to include in the Output Report window and any related export files by going to Configuration lt Output Report Export Fields Select Fields for Qutput Report Bic pics I WW X iv Ping Number I Date Cancel e Time iv Latitude Iv Longitude iv Bottom depth iv Bottom type Iv ED MEI MEZ WET iv Sediment thickness
91. ion Method B1 is generally considered to be obsolete BS amp E 2005 03 0022 2 2 30 4 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Bottom Classification Methods Table 4 2 Summary of Bottom Classification Methods Method Echo Signal Feature Means of Comparison IE E Cumulative Energy o the Least Squared Error Bs all al Energy Ratio E1 E2 0 B Cluster Analysis Second Part of il Bottom id Echo and the Second Bottom Echo 3 7 BED Energy Ratio E1 E1 of the Cluster Analysis First and Second Parts of the First Bottom Echo 2 B4 Ratio of the Fractal Dimension Fractal Dimension of the envelope of the first amp bottom echo energy E1 FD 5 9 10 Cluster Analysis First Bottom First Part E1 The first part of the first bottom echo from the start to peak El typically corresponds to the hardness of the bottom First Bottom Second Part E1 The second part of the first bottom echo from the peak to the end El typically corresponds to the roughness of the bottom Second Bottom E2 The second echo from the bottom resulting from the same ping as the first echo from the bottom 4 2 1 Method B1 First Echo Normalization The First Echo Normalization B1 method developed by Pouliquen and Lurton 3 is generally considered to be obsolete An explanation of the method is provided for historical purposes The First Echo Normalization method was develop
92. ion lt Output Report Export Fields If you make changes to the output report settings via the CONFIGURATION MENU while the analysis is in progress a new output report window will be opened and the remaining data will be displayed in both report windows If you double click on a ping in the output report window it will move to that ping in the oscilloscope window You can export the output report window by the using Copy to Clipboard from the EXPORT DATA MENU or selecting the Copy icon from the toolbar The information can then be pasted into a spreadsheet 3 5 Map window VBT allows you to import a map image and label bottom types directly onto the map A map image is associated with a file by selecting Show Map Window from the EDIT Map Menu or CTRL M You can select an image to use as a map by going to the EDIT MAP MENU and selecting Import Image From File or Paste Image From Clipboard If a map image has already been associated with the file it will open automatically when you select Show Map window When you first import a map file you must define Latitude ux 4 the coordinates of the map Cancel The map coordinates are Longitude Longitude set by going to Edit Map ER di lt Calibrate Map Spatial 2 Range and defining the minimum and maximum lat Latitude BS amp E 2005 03 0022 2 2 30 66 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT Us
93. l University of Gdansk B 1 Introduction Information retrieved from sea bottom echo can be applied in fisheries hydrog raphy marine engineering environmen tal sciences and other fields Acoustic methods used to identify the bottom and bottom sediments are the subject of ex tensive research Attempts to create uni form classification systems for various shelf seas and ocean areas do not yield the expected results The vast variety of types of sediments as well as various ge omorphologic forms and the layer struc ture make it difficult to determine the types of sediments with acoustic meth ods The correlation between physical and statistical parameters of the bot tom echo and the types of sediments are neither unique nor explicit in terms of the majority of the various areas of the World Ocean The acoustic methods of sea bottom classification encompass fol lowing approaches e measurement of energy ratio in the first and second echo reflected from lgprrogs NOTE The editor CJC acknowledges and regrets that an explicit explanation of Method B2 is not included within this text Appendix B B2 amp B3 Publication the bottom 2 3 comparison of the shape of cumu lative averaged envelope of a bot tom echo with patterns obtained from theoretical models of bottom reverberations 5 comparison of measured parame ters of the echo envelope to pa rameters calculated from the the oretical models of r
94. lost bottom signal see Chapter 4 1 View lt Navigate Pings This section of the VrEw MENU contains commands which allow you to move through the pings contained in the data file using the oscilloscope window Go to Ping Number allows you to specify a ping to view Go to First Ping brings the Oscilloscope window to the first ping Use this command to reanalyze a data file after you have made any changes to the settings Go to Previous Ping allows you to view the previous ping in the Oscilloscope window Go to Next Ping allows you to view the next ping in the Oscillo scope window View gt Zoom Zoom In allows you to zoom in in the echogram Oscilloscope or map windows only Zoom Out allows you to zoom out in the echogram Oscilloscope or map windows only View gt Show Map Window The Show Map Window command displays the map window associated with the open data file To associate a map image file with a data file use either Import Map Image from File or Paste Map Image from File from the EDIT MAP MENU and select the map image you wish to use Once a picture file is associated with the data file it can always be opened using the Show Map Window command BS amp E 2005 03 0022 2 2 30 20 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT D BioSonics VBT User s Guide 1 7 3 Configuration Menu The CONFIGURATION drop down menu allows the user to view and or change the defa
95. m four types of the bottom rock BS amp E 2005 03 0022 2 2 30 D BioSonics Results sand mud and mud with sand Since most of the data for 420 kHz did not con tain second bottom echo it was not pos sible to compare the first bottom method and the first second bottom method For this data set the figures illustrate the re sults for two pulse lengths 7 0 2 ms and 0 4 ms Fig 7 In order to visualize the influence of pulses reflected obliquely from the bottom for each series of measured data the results were processed for all pulses For clear visualization of processed data the pulses were averaged over a specific number of consecutive pulses e g 20 pulses The averaged data is presented in the figures The results of the classification are pre sented as points on scattering diagrams Figs 5 6 and 7 93 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 313 kHz gt hardress signature El oo 0 1 1 d 4 5 B T 8 roughness si gratue El b3 120 kHe Ei hardness signature El cn mo 0 1 1 3 4 5 B T 8 roughness signature El 38 kH hardress signature El cn oo 0 1 1 3 4 5 6 T 8 roughness si gratue El dj 120 kH hardness signature El 0 1 3 4 5 B T 8 roughness signature El Figure B 5 Classification results for first echo division method a b pulses with maxi mum am
96. menu Start gt Program Files gt BioSonics gt Visual Bottom Typer VBT as well as the installation directory C BioSonics VBT Please go to C Windows and create a copy of the file VBT ini Name the copy VBT_default ini and store it in the same directory This will allow you to restore the default VBT settings in the future A copy of the file is included in Appendix D PRE RELEASE DRAFT xvili INSTALLATION INSTRUCTIONS PRE RELEASE DRAFT B1 B2 B3 B4 E1 E1 E2 FD NMEA TVG VBT Acronym Guide Bottom Classification Method 1 First Echo Normalization Bottom Classification Method 2 First Second Bottom Ratio E1 E2 Bottom Classification Method 3 First Echo Division E1 EI Bottom Classification Method 4 Fractal Dimension E1 FD Second Part of First Bottom Echo First Part of First Bottom Echo Second Bottom Echo Fractal Dimension of echo envelope National Marine Electronics Association format for longitude and latitude Bottom Classification Method Sediment Time Varied Gain 0 20 log R 40 log R BioSonics Inc VBT SeaBed Classifier X1X PRE RELEASE DRAFT XX PRE RELEASE DRAFT ACRONYM GUIDE Chapter 1 Getting Started 1 1 Introduction BioSonics Inc VBT SeaBed Classifier can be used to determine the composition of the seabed or any other body of water natural or artificial Bottom typing is accomplished through analysis of the bottom echo
97. n Figure 4 2 3 shows the model echo envelope generated by a soft bottom and Figure 4 2 3 shows the model echo envelope generated by different types of the bottom fine sand sand gravel and rock Method B3 calculates the energy of El and El for each ping in the unclassified data set and plots the pings accordingly where El is the vertical axis and El is the horizontal axis The bottom type of each ping is determined by the placement of the feature space boxes previously generated from the ground truth library BS amp E 2005 03 0022 2 2 30 8 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Bottom Classification Methods Figure 4 7 Results of echo envelope simulation for soft mud using an ideal cone beam pattern of 26 width bottom surface reverberation blue bot tom volume reverberation green sum red 18884 Figure 4 8 Results of echo envelope simulation for different types of the bottom fine sand red sand blue gravel green and rock grey taken using a 26 transducer 4 2 4 Method B4 Fractal Dimension Method B4 classifies the bottom type by characterizing the shape of the bottom echo It is known that the shape and structure of a bottom echo curve is unique to the bottom type However the curve is a complex structure that cannot be simply defined in Euclidean geometry Fractal dimensions can be used to describe complex shapes
98. n of se bottom type Ocenolo gia 19 61 78 4 Stepnowski A Bakiera D Moszyn ski M 1996 Analysis and simulation of hydroacoustic methods of sea bed clas sification raport badawczy 52 95 Po litechnika Gdanska Gdansk 5 E Pouliquen and X Lourton 1992 sea bed classification using echo sounder signals Proceedings of European Con ference on Underwater Acoustics Lux BS amp E 2005 03 0022 2 2 30 Ph D thesis 99 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 ChapterB Methods B2 amp B3 BS amp E 2005 03 0022 2 2 30 100 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Appendix C Method B4 Publication Sea bottom typing using fractal dimensions vspace lin Z Lubniewski amp A Stepnowski Technical University of Gdansk Acoustics Department 80 952 Gdansk Poland Paper published in Proceedings of the International Symposium on Hydroacoustics and Ultrasonics Gdansk Jurata Poland 12 16 May 1997 C 1 Summary The article presents an attempt to ap ply elements of fractal analysis for the purpose of sea bottom typing The frac tal dimension was calculated as box di mension for sampled envelopes of echo signals from four types of sea bottom recorded during mobile acoustic surveys carried out in Lake Washington The results obtained show that the simple method applied can be used for on board sea bed
99. n the CONFIGURATION of the sampling MENU items windows not their location along the Step 1 Adjust the thresholds horizontal lines O EN Step 1a The lowest threshold Data Processing Filter Threshold should be greater than or equal to the Data Acquisition Threshold found in Configuration lt Data File Trans ducer Properties Step 1 b The middle threshold Bottom Detection Threshold can be numerically defined by going to Configuration lt Bottom Tracking Parameters The Bottom Detection T hresh old should be a few decibels higher than the end of the bottom echo Step 1 6 The top threshold Peak Threshold can be numerically defined by going to Configuration lt Bottom Tracking BS amp E 2005 03 0022 2 2 30 49 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter2 Tutorials Step 2 Step 3 Bottom Typing B1 82 p3 84 s Parameters The Peak Threshold should be a few deci bels below the highest peak in the data file In the Oscilloscope window the sampling windows are represented by vertical dotted lines You can manually adjust the width of the sampling windows by dragging right hand boundary of the window Dragging the left hand boundary of the window allows the user to shift both sides of the window simultaneously while maintaining the width of the window Clicking on the left hand boundary of El the sta
100. nc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide the analysis of the experimental data Such a selection of integration bound aries will assure elimination of influence of volume reverberation component for its roughness when this level is low Fig 3 In order to eliminate the influence of the pulses scattered at oblique inci dence the echoes reflected at normal in cidence should be extracted from the se ries of data A simple method of clas sification was used based on extracting pulses that were presented as local max imums in the series of pulses To show the effect of obliquely scattered pulses the results of the analysis of the whole amount of data are presented To retain the readability of the diagrams the results of the total series analysis were averaged in the groups of equal size The averaged results are presented in the figures B 3 Results Analysis was performed for hydroacoustic signals from four types of the bottom rock sand soft sand and soft mud The frequency of the echosounder was 38 kHz and 120 kHz Two series of data were analyzed for every class of the bottom one for each frequency Fig 5 For comparison classification with another method was also performed The other method classifies the signal using an in tegral of the second echo envelope as a measure of bottom hardness Fig 6 E2 V t dt 0 For the 420 kHz data bottom signals were fro
101. nc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide only by special range adjustment in the echosounder In most of the cases deep seas soft bottom types a high dynam ics echosounder is required for that pur pose In relation to other methods of sea bot BioSonics References embourg Sept 1992 535 538 6 Tegowski J 1994 Charakterysty czne cechy rozpraszania wstecznego syg nalw ultradzwiekowych od dna w Bal tyku Poludniowym Doktorat I O PAN Sopot Characteristic features of back scattering tom classification using for example neuralof acoustic signals in South Baltic Sea networks or cluster analysis presented method characterizes with a big com puting simplicity It enables the method to be implemented in a portable system on board of the research vessel work ing in the real time The advantage of the method is also the fact that it has no time consuming learning phase and there is no need to use a great amount of data for it B 5 References 1 Bakiera D 1995 Analiza numeryczna sygnalu echa hydroakustycznego od dna morskiego praca dyplomowa Politech nika Gdanska Gdansk Numerical analy sis of acoustic echo signal from the bot tom Master thesis 2 Chivers R C Emerson N Burns D R 1990 New acoustic processing for underway surveying The Hydrographic Journal 56 8 1T 3 Orlowski A 1984 Application of multiple echoes energy measurements for evaluatio
102. nd structure of naturally methods e g geological cores or remotely occurring structures cannot usually be operated vehicles with TV cameras as being non invasive more cost effective and faster In general the following ap defined in terms of Euclidean geometry That is why this kind of geometry is not the most adequate tool to describe this 101 PRE RELEASE DRAFT VBT User s Guide type of elements On many occasions however nature has proved to accom modate various types of elements with fractal structure e g the structure of plants leaves corrugated sea surface or bottom surface 3 which suggests that fractal analysis methods are the proper methods to study and describe such el ements Fractal sets are defined as scale invariant self similar geometric objects A geo metric object is called scale invariant if it can be written as a union of rescaled copies of itself Regular fractals such as the Cantor set Sierpiski triangle or Koch snowflake which is shown in Fig 1 display exact self similarity 3 4 Random fractals display a weaker sta tistical version of self similarity When we want to measure the size of fractal figures we encounter problems We are not able to measure the size of such figures using standard methods The Koch Snowflake for instance has a sur face equal to zero we mean the side of this figure but its length tends to infin ity when the size of the measur
103. nd truth data file was opened where the first color in the Custom Color section corresponds to the first file opened see 22132 BS amp E 2005 03 0022 2 2 30 24 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Menubar Commands Configuration gt Bottom Sampling Windows The Bottom Sampling Windows Bottom Sampling Windows BioSonics Inc xx menu items allow the user to 5 specify the widths of sam ERN pling windows for the follow E1 First Bottom Second Part jas samp Cancel ing events first bottom first E2 Second Bottom Window samp part E1 first bottom second sediment Window C part E1 second bottom win dow E2 and sediment win dow S The Data acquisition pulse duration is provided for reference and cannot be altered Data Acquisition Pulse Duration 5 samp We recommend using the following values See Chapter 4 1 for more infor mation on bottom tracking Table 1 3 Recommended Bottom Sampling Window Widths Sampling Window Methods B2 B3 amp B4 Method B1 Configuration gt Bottom Tracking Parameters The Bottom Tracking Pa rameters Window allows the user to improve the ac lll Bottom Tracking Parameters BioSonics Inc Bottom Detection Peak Threshold 53 148 Cancel curacy of the bottom detec oe tame tion algorithm by manually Bottom Detection Threshold dB specif
104. nformation on working with the method windows BS amp E 2005 03 0022 2 2 30 6 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Using VBT An Overview First Echo Normalization B1 T Info gt gt Max AMS 007000 Info gt gt Error 007000 RA x 7 DIETE Import info gt gt E qoa Close Window EO 1 18202e 005 Thickness Figure 1 5 The five method windows available in VBT BS amp E 2005 03 0022 2 2 30 7 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started Table 1 2 Sample library included with VBT Filename File type Sampling Bottom Type Frequency mud mud TRIBU IokH al File Type TR7B2 TXT output report from TR7_B dt4 analyzed with method B2 TR7B3 TXT output report from TR7_B dt4 analyzed with method B3 TR7B4 TXT output report from TR7_B dt4 analyzed with method B4 B_234 XLS comparison of bottom classification results using methods B2 B3 and B4 on transect TR7_B dt4 1 4 Sample Library VBT requires that the user provide a reference catalog of physically verified data files referred to as the ground truth library The VBT software package includes a small reference catalog containing ground truth data files for sand rock and mud collected in Liberty Bay near Seattle Washing
105. o 222 58 0 4 120000 1490 34 C ABIOSONICSAVBTACTLGAMUD120 DT4 57 69 50 16 48 100 0 23870 55 9157 o 222 58 0 4 120000 1490 34 C BIOSONICS VBT CTLG TR7_B DT4 53 69 50 16 48 100 0 23870 7 1 222 58 0 4 120000 1490 34 SAND120 DT4 53 69 50 16 48 100 0 23870 55 9157 5 222 58 0 4 120000 1490 34 ROCK120 DT4 40 69 50 16 48 100 0 23870 55 9157 5 222 58 0 4 120000 1490 34 TR7_B DT4 53 69 50 16 48 100 0 23870 55 9157 1 222 58 0 4 120000 1490 34 MUD120 DT4 53 69 50 16 48 100 0 23870 55 9157 5 222 58 0 4 120000 1490 34 Window size Rect 0085 0064 0712 0483 icon 0 max 1 tool 1 status 1 Recent File List File1 C NBIOSONICSNVBTNCTLGNMUD120 DTA File2 C BIOSONICS VBT CTLG SAND120 DT4 File3 C BIOSONICS VBT CTLG ROCK120 DT4 File4 C BIOSONICS VBT CTLG TR7_B DT4 Method 3 E1Max 10 E2Max 10 E1Min 50 E12Min 50 BS amp E 2005 03 0022 2 2 30 110 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 LinLog 1 Rect5Library01 No 4 Rect1 62 70 97 109 Rect2 94 45 130 74 Rect3 41 13 73 34 Rect4 115 31 147 54 Info00 120 kHz Info01 FBW30 60 20p Info02 red SMUD green SSAND blue HASAND pprp ROCK Info03 leaveonly SSAND Library01 No 2 Ping10 0 0049 0 0141 0 0273 0 0461 0 0745 0 1166 0 1752 0 2523 0 3436 0 4365 Ping11 0 5205 0 5903 0 6492 0 7059 0 7606 0 8099 0 8521 0 8869 0 9185 0 9485 Ping12 0 9721 0 9867 0 9951 1 0000 Info0
106. on from the bottom surface Its amplitude is determined by the ratio of acoustic impedance of the sea water and the bottom material As a result it was BS amp E 2005 03 0022 2 2 30 El V t dt 0 The attack phase starts at the start of the echo signal and its duration equals the sounding pulse length During the decay and release phases the echosounder transducer receives signals scattered at oblique incidence on the bot tom surface The higher the bottom rough ness the higher the amplitude of this component 2 Therefore the integral of the echo envelope in the range of the decay and release phases is treated as a measure of bottom roughness Vmax 60dB El J Vi t dt The threshold 60 dB in the above for mula is an arbitrary threshold that suf ficiently minimizes the disturbance level for estimated parameter During these two phases an increasingly stronger echo signal originating from volume rever berations in the bottom interferes with the bottom typing process As such the integration interval should be selected in a way that will minimize these ef fects In order to do this the start of the decay phase should be located in the envelope maximum and its end at the point when the decaying slope of the echo decreases to 60 dB level corre sponding to the moment the pulse com pletely penetrates the bottom 1 The value of the threshold is the result of 92 Copyright C 1998 2005 BioSonics I
107. onds to an untyped bottom 2 42 Output Report Window In the output report window the bottom typing results are displayed nu merically in the Type column where the number corresponds to the order in which the feature space boxes were created For example in the tutorial in Sections 2 1 1 amp 2 1 2 mud120 dt4 was the first file opened and therefore each ping that was classified as mud will have a 1 in the Type Column A zero is displayed when the bottom was not typed BS amp E 2005 03 0022 2 2 30 52 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Interpreting the Analysis Results 2 43 Map Window If a the map window was open during analysis the bottom typing results will be plotted on the map as a color coded dot at each ping As in the echogram window the color code corresponds to the colors established in the method window A white dot corresponds to an untyped bottom 2 4 4 Improving the Accuracy of the Analysis Lost Bottom The first step in the VBT bottom typing algorithm is to isolate the bottom echo portion of each ping bottom tracking If the bottom tracking parame ters are not properly set the algorithm may miss may of the bottom echoes lhe bottom tracking parameters can be set by going to Configuration gt Bottom Tracking Parameters You can determine the number of lost bottoms in your analysis by right clicking on the echogram window Ea
108. ope window for file sand120 dt4 The color shown here may differ from the colors on your screen However the data shape of the signal should be identical You can customize the colors your oscilloscope window by going to Configuration gt Colors Step 2 Preparing for analysis Step 2a The group of numbers in the lower right hand area of VBT corresponds to the location of the mouse cursor in the os cilloscope window Sample Number the Depth the cursor is on in E cursor is onin the the oscioscope window Data Acquisition oscilloscope window Pulse Duration 11 140 1224 70 00 26 59m 120kHz 4ms Charnel Number Active Ping Total intently Value the Data Acquisition Number ofP ings is on in the Samblina Frequency pne Window The same informaion is displayed in the heading of the oscioscope Window Step 2b In the oscilloscope window use the mouse to change the thresholds horizontal lines See Section 3 1 for more infor mation on manipulating the oscilloscope window and Sec tion 4 1 for more information on the bottom tracking para meters Step 2b 1 The lowest threshold Data Processing Filter Thresh old should be greater than or equal to the Data Acquisition Threshold Step 2b 2 The middle threshold Bottom Detection Thresh old can be numerically defined by going to Configuration gt Bottom Tracking Parameters Step 2b 3 The top threshold Peak Threshold can be nu merically defined b
109. or the oscilloscope window Ping allows you to select the color of the signal plot in the oscilloscope window First Bottom Windows allows you to select the color of the sam pling windows vertical lines for the first bottom echo El and E1 Second Bottom Window allows you to select the color of the sam pling window vertical lines for the second bottom echo E2 Thresholds allows you to select the color of the bottom tracking thresholds peak threshold bottom detection threshold and data analy sis threshold Echogram allows you to customize the echo intensity color bar right hand side of the echogram window Each color can be manually se lected by first choosing a color in either Basic Color or the color chart and then selecting Add to Custom Colors The colorbar col ors will be displayed in the order of the colors in the Custom Color section Color E 2 x Basic colors E mE Custom colors Hue 160 Ta a lT MN Greer 0 Hn stam E gt ColorlS olid Lum o Blue o STE Cancel T Add to Custom Colors 60 Bottom Types allows you to customize the color code for each bot tom type for the echogram window and the method windows Each color can be manually selected by first choosing a color in either Basic Color or the color chart and then selecting Add to Custom Colors lhe color code for each bottom type is determined by the order in which the grou
110. orizontal axis displays the color coded results of the bottom classification where the colors are defined in the method windows BS amp E 2005 03 0022 2 2 30 4 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Using VBT An Overview Echo Intensity Color Bar dE Bottorn Classification Color Bar Figure 1 2 An example of the echogram window from sample library file TR7_B dt4 See Section 3 3 for more information on working with the echogram window The output report window Figure 1 3 displays a table of the analysis results STR B DT4 2 E nm xj Ping Date Time Latitude Longitude Depth Type EU El ES 871 18 11 95 02 02 12 00 547 42 75488088H 122 37 890000 7 87 5L H5e H Bd 6 4 75e 003 1 64 891 18 11 95 amp 82 82 32 88 57 542 758B88H 122737 90800004 7 88 A 5 885 886 3 81e 883 10 11 95 82 82 52 88 47 52 7688880H 122 S37 910000 7 46 A 5L HB8e H B 3 82e 883 1 1 931 18 11 95 82 83 12 B88 47 42 81000600H 122737 940000 7 71 8 5 885 886 2 91e 803 1 1 18 11 95 62 03 32 00 27782 82888878 122737 0000W 8 5 886 886 1 47e 003 1 Figure 1 3 An example of the output report window from sample library file TR B dt4 The output report window is organized by ping number and displays the data numerically for the following fields Date date the data was collected Time time the ping was recorded Latitude l
111. oving the Accuracy of the Analysis 2 5 Exporting the Analysis Results 2 6 Analyzing Batch Files 2 7 GIS Ready Analysis 2 2 4 4 442 244 4 od 3 VBT Windows 3 1 Oscilloscope Window 3 1 1 Analysis Thresholds 3 1 2 Sampling Windows Slo A ate aru e or d AE Asp CO kw E ARE RM euo 3 2 Method Windows wor cto RSS 3 2 1 Components of the Method Window 023 Methods B2 B9 BA a tae ne amp S 3 2 3 Fuzzy C Means FCM Clustering 3 2 4 Manipulating the feature space boxes 3 3 Echogram Window 3 4 Output Report Window 9 0 Map 0 4 p e 0 5 9 XXe 8 8 0 SAR 4 How VBT Works 000 Tracking 608 amp a 4 2 Bottom Classification Methods 4 2 1 Method B1 First Echo Normalization 4 2 2 Method B2 First Second Bottom Ratio 4 2 3 Method B3 First Echo Division 4 2 4 Method B4 Fractal Dimension 4 2 5 Method S Sediment Layer Appendices A Method B1 A l Introduction A2 Reula atico dea a A3 DISCUSSION s eu 42 gt a od oe AS AA RETENE 4 42 gt 8 2 xe mo 1 SR X 4 CS PRE RELEASE DRAFT CONTENTS CONTENTS B Methods B2 amp B3 Bek Introduction 456 oed ext x 9 Ed B 2 First Echo Division Bo RoUe so d ode 4 2 6
112. plitude c d averaged for all pulses rock sand o soft sand x soft mud BS amp E 2005 03 0022 2 2 30 D BioSonics ChapterB Methods B2 amp 3 El Eh h hardness agrature E E th hardness signature El roughness signature El a 33 LHe Po sir ma gras m E LgEgR LR F dj 120 kH hardress signature El o 1 13 4 6 8B 7 8 roughness signature El Figure B 6 Classification results for the second echo method a b pulses with maxi mum amplitude c d averaged for all pulses rock sand o soft sand x soft mud 94 Copyright 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT BioSonics VBT User s Guide A Results a jT 0 2 me hardness signature E1 E DO A 0 5 0 1 0 15 02 D 25 roughness si grate El bj TEO Ame hardness signature El 0 0 05 Of 015 02 025 Of 035 348 roughness signature El c T20 2 me ig oor E cn hardness signature E1 E E D oz 0 002 004 D 06 8 0 1 0 12 roughness signature El dj Te0 4 ms hardness signature El 0 007 004 006 0 08 Di 012 014 roughness signature El Figure B 7 Classification re sults for the second echo method frequency 420 kHz a b pulses with maximum amplitude c d averaged for all pulses rock sand o soft sand x soft mud BS amp E 2005 03 0022 2 2 30 95 Copyright C 1998 2005 BioSonics Inc All Rights Reserved
113. process by going to Configuration gt Oscilloscope and selecting the corresponding channel If you wish to process another channel select go to first ping in the VIEW MENU or toolbar and reanalyze the data Be sure to review the settings in the oscilloscope window See Section 2 2 1 New echogram and output re port windows will open Bl Oscilloscope Options BioSonics Inc AX TIWh None 20LogH C 40LogR Beam Type 4 Narrow Cuide Cancel Channel lea Cm C3 Ca Ch CB OF 22 Depth x Scale Max C Samples F Meters A Mees p poen Initial Offset Samples Per Meter 26 499 cAmplitude 1 Scale Log Cim Background Color Fixed Solid Color Chart Type G Line Samples C Echogram Colors 2 4 Interpreting the Analysis Results 2 4 1 Echogram Window In the echogram window the bottom typing results are displayed on the bottom axis A vertical colored line is present for each bottom typed ping The color of the line corresponds to the color of the feature space window for the bottom type as defined in the method window For example in the tutorial in Sections 2 1 1 amp 2 1 2 mud120 dt4 was the first file opened and therefore has a red feature space box in the method window In the echogram window pings that were classified as mud have a red line in the bottom axis A white line corresp
114. pt d ns MA Each line of the resulting text tone file is a single ping and each number contained in the line is a sample yw Note The Data Acquisition Threshold found in Configuration gt Data File Transducer Properties is displayed for any sample whose in tensity is less than or equal to the data acquisition threshold 8This function does not work properly for this release of VBT BS amp E 2005 03 0022 2 2 30 30 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Menubar Commands Start Ping allows you to specify the first ping to include in the data file End Ping allows you to specify the last ping to include in the data file The last ping is displayed by default Start Sample allows you to specify the first sample to include in the data file End Sample allows you to specify the last sample to include in the data file y Note Specifying a sample range here will export the same range for all pings Export all samples to avoid losing part of your data file Filename allows you to specify the path for the output file Percent Complete displays the status of the exportation OK select OK to export the pings to file Cancel select Cancel in order to cancel the export Export Ping Options BioSonics inc x C Entire ping Selected area Cancel Bottom areas only Options allows you to e
115. recognition in real time with ac curacy similar to that of other methods Introduction C 2 Sea bottom identification methods have a wide range of applications in hydrog raphy marine engineering environmen tal sciences fisheries and other domains Acoustic methods which use the infor mation retrieved from the acoustic bot tom echo have advantages over the other proaches are used in the acoustic meth ods of bottom typing e measurement of energy ratio of the first and second bottom echo so called RoxAnn method 2 e comparison of the actual cumula tive echo envelopes with theoreti cal patterns 5 e analysis of a set of values of acoustic and statistical parameters of the echo envelope using cluster analy sis or neural networks 6 7 e division of the first echo signal 1 It is known that the surface of sea bot tom is an example of a naturally occurring fractal structure 3 Taking this into con sideration as well as the fact that frac tal dimension is a measure of complex ity the authors made an attempt to use fractal dimension of the recorded echo envelopes in order to perform sea bottom typing The implicit assumption here was that fractal structure of the bottom is transferred onto sort of its image ob served in the form of an envelope of a sonar echo C 3 Fractal analysis In the course of examining and describ ing the elements of nature it was found that the shape a
116. ring algorithm for automatic classification of fields for different bottom categories in x y coordinates for methods B2 B3 B4 Fuzzy C Means FCM is a data clustering technique wherein each data point belongs to a cluster to some degree that is specified by a membership grade This technique was originally introduced by Jim Bezdek 1981 as an improvement on earlier clustering methods It provides a method to group data points that populate some multidimensional space into a specific number of different clusters FCM algorithm starts with an initial guess for the cluster centers which are intended to mark the mean location of each cluster The initial guess for these cluster centers is most likely incorrect Additionally FCM assigns every data point a membership grade for each cluster By iteratively updating the cluster centers and the membership grades for each data point FCM iteratively moves the cluster centers to the right location within a data set This iteration is based on minimizing an objective function that represents the distance from any given data point to a cluster center weighted by that data point s membership grade Figure 4 2 5 shows clustering of Fractal Dimension data E lt Fuzzy C Means Clustering Number of clusters 3 Database option paa Display cluster centers only sand 20 ct4 2 0 ala C Create new database gt E e a Append to current database pes Fractal Dimension Metho
117. rt Step 15 Step 16 Bl Fractal Dimension B4 Method Biosa Minimize all windows Open file sand120 dt4 and repeat Steps 6 13 y Note In the method window the box will be green The color of the box is determined by the order in which the box was drawn The significance of the color of the dots does not change in Report View Right click on the titlebar in the method window and select Report View In Report View the method window should look similar to the following figure gt x Info gt gt ED import Export Right click on the titlebar in the method window and select Multifile View Multifile View shows the average dots for each open file The color of the dots corresponds to the color of the feature space box e g red for mud and green for sand BS amp E 2005 03 0022 2 2 30 46 Copyright 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Ground truth Library Fractal Dimension B4 Method BioSonic Step 17 Minimize all windows Step 18 Open file rock120 dt4 and repeat Steps 6 13 y Note The feature space box will be blue for the third file rock120 dt4 Step 19 In Report View the feature space should look as follows y Note the sand feature space green box was adjusted to avoid overlap with the rock feature space blue box Step 20 Info gt gt FD 1 10001 mm bel Step 21 Minimize all w
118. rt of the bottom echo will allow you to shift all windows simulta neously while maintaining their width and relative positions Adjust the sampling windows vertical lines The sample windows should be adjusted so that the E1 sampling win dow the second half of the first bottom echo has red hatch marks in the oscilloscope window You can manually adjust the width of the sampling windows by dragging right hand boundary of the window Dragging the left hand boundary of the window allows the user to shift both sides of the window simultaneously while maintaining the width of the window Clicking on the left hand boundary 01 6 start of the bottom echo will allow you to shift all windows simulta neously while maintaining their width and relative positions Refer to Section 93 1 for more information on working with the oscilloscope window 2 2 2 Echogram and Output Report Windows Settings If you wish to change any of the default settings for the echogram window and output report window NOT RECOMMENDED you must make your selections prior to analyzing the data file 2 2 3 Bottom Typing Method Once you have set up the bottom tracking parameters and bottom sampling windows in the oscilloscope windows you may select the bottom typing method you wish to use Select Bottom Typing Method i Press the corresponding method button in the toolbar BS amp E 2005 03 0022 2 2 30 50 Copyright C 1998 2005 BioSonics Inc All Rights R
119. sented investiga tion are promising as they show that evaluation of fractal dimension may be a usable method of bottom typing The results are not worse than those obtained concurrently using other methods How ever it must be noticed that the data used were restricted only to one rela tively small water region so the method should be verified for a larger area and for different parameters of data acquisi tion In this work the authors assumed that fractal structure of the bottom sur face transfers itself onto its image in the echo envelope It is the authors opin ion however that an extension of the applied method could be the applica tion of the deconvolution of scattering impulse response of seabed from bottom echo before the fractal dimension is cal culated C 7 References BioSonics ChapterC Method B4 ter Acoustics Elsevier Applied Science London and New York 535 6 Stepnowski A Moszynski M Komen darczyk R Burczynski J 1996 Visual real time Bottom Typing System VBTS and neural networks experiment for sea bed classification Proceedings of the 3rd European Conference on Underwater Acoustics Heraklion Crete 685 690 7 Tegowski J 1994 Characteristic fea tures of backscattering of the ultrasonic signals from the sea bottom at the South ern Baltic in Polish Ph D Thesis In stitute of Oceanology of Polish Academy of Sciences Sopot 1 Bakiera D Stepnowski A 1996 Method
120. such as an energy curve by treating the complex shape as a collection of simple geometric shapes The fractal dimension FD of a geometric shape is equal to log N log S 4 1 where N is the number of self similar parts contained within the shape and S is magnification factor required to make one self similar part the size of the original shape BS amp E 2005 03 0022 2 2 30 79 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter4 How VBT Works For example an equilateral triangle can be divided into four equally sized triangles N is four Figure 4 9 Equilateral triangle containing four self similar equilateral trian gles A self similar part needs to be magnified by two to equal the size of the original shape S is two Figure 4 10 A self similar triangle from Figure 4 2 4 B Original triangle in Figure 4 2 4 Magnifying the self similar triangle by two results in the original triangle Since the value of aes is equal to two the fractal dimension of the triangle is two which is also equal to its Euclidean dimension Method B4 in VBT performs a similar calculation on the energy curve of E2 the second part of the first bottom echo 4 2 5 Method S Sediment Layer The time integral of echo signal envelope is calculated for this depth layer The user can apply the echo integration model without attenuation or with attenuation For
121. ter Centers Only displays the cluster circles in the all databases and does not display corresponding feature space boxes Create New Database creates clusters and feature space boxes in a new empty database The new empty database must first be created by selecting New in the Method window Append Current Database Adds a cluster to the existing data base A cluster will be added for any data file that is open Rectangle Size Allows you to specify the size of the cluster space A multiplier of one will use dimensions of one cluster variance A mul tiplier of three will use dimensions of three times the cluster variance Results Filename datafile where the coordinates of the cluster are stored Be sure to change the file name or the default filename will be used and you will lose your previous data BS amp E 2005 03 0022 2 2 30 61 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter3 VBT Windows 3 2 4 Manipulating the feature space boxes The boxes in the method window can be adjusted using your mouse The cross hatch cursor allows you to move the box while maintaining its shape and the diagonal cursor allows you to resize the box You may delete a tem plate by right clicking on it Double clicking on any template will highlight all templates BS amp E 2005 03 0022 2 2 30 62 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RE
122. the signal with attenuation the user should find extinc tion and attenuation parameters by experiments most likely sigma back scattering per kg and sigma extinction per kg Back scattering and extinc tion parameters can be entered by a user into Parameters dialog box using menu command Options Transducer We expect that this layer can be used for quantitative estimate of aquatic plants from bottom fish and also from sediments A backtracking bottom tracking method is used in the VBT software This means that the software begins to track the bottom from the far end of the received echo and moves toward a shorter range Then the BS amp E 2005 03 0022 2 2 30 80 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Bottom Classification Methods bottom tracker locks to the rising edge of the first bottom echo The sedi ment layer is at the range before the bottom echo i e it is the layer above the bottom The sampling gate called in VBT sediment layer is open for sampling either sediment or bottom plant echoes The VBT software is estimating the thickness of this layer and also the integral of the echo BS amp E 2005 03 0022 2 2 30 81 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter4 How VBT Works Fuzzy C Means FCM Clustering The VBT software version 1 5 includes the FCM Fuzzy C Mean Cluste
123. thod compares a portion of the bottom echo signal from the un known data set with the same portion of the bottom echo signal from the ground truth data set In Method B1 least squared error calculations are used to classify the unknown data set by comparing the energy curve of each ping to energy curves for each bottom type included in the ground truth data Methods B2 B3 and B4 use cluster analysis techniques to classify the unknown data set In these methods the feature space is divided into regions for each bottom type e g rock sand or mud and the bottom type for each ping is determined by the feature space it falls into Table 4 2 provides a summary of the bottom classification methods available in VBT The information about hardness and roughness of the bottom is encoded in echo signal envelope The signal is decoded by taking the time integral of bottom echoes and extraction of hardness and roughness signature The hardness and roughness signatures can be projected in X Y coordinates and different types of the bottom can be classified by determining different areas and boundaries of ROUGHNESS HARDNESS signatures values correspond ing to different bottom type e g mud sand etc y Note There are different axes for each of the analysis methods Method B1 plots the analysis results as energy versus maximum root mean square RMS error Methods B2 B3 and B4 plot the analysis results with one axis for each feature used for classificat
124. to a single named user 3 Institutional Multi Seat License In consideration of payment of the Institutional multi seat license fee which is part of the price paid for this software package referred to in this Agree ment as the Software BioSonics Inc as Licensor grants to you as Li censee a non exclusive right to use and display this copy of the Software on as many single computers i e single CPUs as you have purchased license seats for and to assign their use to an equal number of named users affiliated with your Institution For example purchasing three 3 license seats entitles you to use and display this Software at any time on three computers i e three CPUs and to allow at most any three users from your Institution to be using the Software at any given time PRE RELEASE DRAFT For all licenses described above Additional license codes or transferal of license codes between computers can be arranged by contacting BioSonics Inc support biosonicsinc com Permitted Use of the Software The Software and accompanying written materials are protected under United States and international copyright laws Unauthorized copying of the Soft ware is expressly forbidden You may be held legally responsible for any copyright infringement that is caused or encouraged by your failure to abide by the terms of this Agreement Subject to these restrictions you may make one 1 complete copy of the Softw
125. ton USA using two different acoustic frequencies 120 kHz and 420 kHz Divers physically verified the bottom types A sample transect is also provided in order for you to explore VBT bottom classification methods The complete training set can be requested from 1210502108 Inc lable 1 2 summarizes the characteristics of the reference catalog provided with VBT and stored in C BioSonics AVBT CTLG We recommended that you practice VBT classification with the reference catalog and the sam ple transect before moving analyzing your own data BS amp E 2005 03 0022 2 2 30 8 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Basic Steps 1 5 Basic Steps for Analyzing a Data File in VBT The following introductory tutorial outlines the basic steps for working with a data file VBT and does not define significance of the parameters discussed This tutorial assumes that a ground truth library has been previously estab lished Please refer to the cross referenced sections for in depth explanations and recommendations for each step In order to analyze a file using VBT the following steps need to be taken 1 load the data file 2 prepare the data file for analysis 3 analyze the data file 4 interpret the analysis results and 5 export the analysis results Preparing the data file for analysis is an important step toward accurate bottom typing resul
126. ts Step 1 Loading a Data File Step la Open data file sandD120 D T4 found in the CTLG folder There are three ways to open a data file i Select open file icon in the toolbar p ii Select Open in the FILE MENU mem 111 Press CTRL O with your keyboard Step 1b When you open a data file a pop up window will appear listing the data acquisition parameters for the file Click OK You may refer back to the data acquisition parameters by selecting Data File Transducer Properties in the CONFIGURATION MENU 1 CABIOSONe1TVBTACTLGYSAMD120 DT4 Channel 1 Type Single Number of Samples per Ping 2200 Number oF Pings 140 Transducer Frequency 120kHz Beam Width 6 Pulse Length D ms Step 16 The data file will be displayed graphically in an oscilloscope window shown in the following figure The oscilloscope window plots the intensity in decibels of the echo signal 3Exporting the results is based on the user preference and is not required BS amp E 2005 03 0022 2 2 30 9 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Open File VBT User s Guide 2 Chapterl Getting Started with respect to depth in either sample or meters You will use the oscilloscope window to specify sampling windows and sampling thresholds while preparing the data file for analysis Echo Intensity Axis dB Horizontal Axis Sample Number Figure 1 6 Oscillosc
127. tting Bottom Tracking Parameters 13 Summary of Bottom Classification Methods repeated 75 Percentage contents of particles of different size in bottom SaO 2 ad a a 96 1X PRE RELEASE DRAFT BioSonics Inc VBT SeaBed Classifier End User License Agreement Introduction This is a legal agreement between you and BioSonics Inc The terms of this Agreement govern your use of this software Please read this agreement carefully before you install this package By installing this package you are agreeing to be bound by this Agreement If you do not agree to the terms of this Agreement please return to your retailer the entire package with the original wrapping for a full refund License Grants 1 Trial Use License If you have not purchased a license from BioSonics Inc bio biosonicsinc com you may install the software on any computer for trial use The trial period is 40 days after which you may either purchase a license or request an extension of the trial period support Gbiosonicsinc com 2 Single User License In consideration of payment of the single user license fee which is part of the price paid for this software package referred to in this Agreement as the Software BioSonics Inc as Licensor grants to you as Licensee a non exclusive right to use and display this copy of the Software on a single computer i e a single CPU at one location at any time and to assign its use
128. ttom When a bottom echo is not found within the bottom tracking window it is referred to as a lost bottom BS amp E 2005 03 0022 2 2 30 72 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide A Bottom Tracking Table 4 1 Guidelines for Setting Bottom Tracking Parameters Peak Threshold 15 dB to 30 dB Use a high dB for hard bottoms and a low dB for soft Peak Width 06 m to 1 0 m 09 m Use a narrow width for hard bottoms and a wide width for soft Bottom Detection 30 dB to 50 dB 20 dB less than Use low dB values Threshold Peak Threshold for targets far from the bottom Above Bottom 02 m to 08 m 02 m Use wider settings Blanking Zone for targets far from the bottom Alarm Limit 2 to 20 Use a high number for Tracking Window 1 m to 10 m 2m Use wide settings for Rm o atric BS amp E 2005 03 0022 2 2 30 3 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapter4 How VBT Works 4 2 Bottom Classification Methods When an underwater acoustic pulse is generated information about bottom type particularly its density and texture bottom sediments and aquatic plants is embedded in the echo signal VBT uses signal processing techniques to decode the signal and identify the composition of the seabed bottom VBT provides four bottom typing methods for the user B1 B2 B3 and 4 Each me
129. ucer Properties Ctrl Enter bottom typing algorithm ostitescope Ech Color Bar by highlighting Con CU BB US Colors i fi gur e B ottom Typ ing SetMap Spatial Range Algorithm and the Bottom Sampling Windows Bottom Tracking Parameters appropriate method Wln Select Bottom Typing Method Configure Bottom Typing Algorithms Bi First Echo Normalization Method Ckrl 1 dow The bottom e B2 First 5econd Bottom Echo Ratio t th d 0 Sue B3 First Echo Division Method ctra yping Method 1n use Dy Output Report Filters B4 Fractal Dimension Method Ctrl 4 the algorithm is high pt Report Export Fields gt Sediment Visualization CHS lighted with a red rec tangle Configuration gt Output Report Filters In the Configuration menu Output Report Filters allows E filter X you to select the energy filter ii Pings per Report Cancel M Report the average of all qualifying pings Configure Analysis Output Report BioSonics Inc E Xl and the number of pings per re Stop analysis after one report BS amp E 2005 03 0022 2 2 30 27 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started port Each line in the Output Report Window is one report The value entered in the energy filter specifies the lowest amount of energy a ping can have to be included in the analysis The energ
130. ult analysis settings e g filter threshold or the number of pings per report When changes to the default settings are made BEFORE the data file is played the changes will be implemented to the active file and any future data file If the changes are made AFTER the file has been played the changes will be implemented when the file is replayed or when the original file is closed and any subsequent data file is opened Configuration gt Data File Transducer Properties To view the transducer and data file properties go to Configuration gt Data File Transducer Properties Select Advanced Properties to view all data acquisition parameters You may only vary the following parameters Sampling Rate Sounder Depth m e sound speed Receive Sensitivity dB Data Processing Threshold dB e Data processing filter threshold Data Acquisition Threshold dB Use the Data processing filter thresh old to reduce the effect of noise in the signal present near the Data ac quisition threshold The remaining fields describe the parameters during data acquisition and therefore cannot be changed Full Beam Width Transducer Senal Number 21 35802 0 Advanced parameters Pulse Duration ms Samples Per Meter samples per second Sound Speed m s Source Level dB Filter Ping Rate ms ping Menubar Commands configuration Data File Transducer Properties
131. uted for the ideal beam pattern and a plain homogeneous bottom showed that the certain phases of signal can be dis tinguished in the bottom echo envelope Examples are presented on the Fig 1 showing that distinct phases of an echo signal are independent of the class of the bottom and dependent upon the geo metrical configuration of the model Figure B 1 The results of echo envelope simulations for bottom types fine sand sand gravel and rock for ideal cone beam pattern of a 26 and b 13 width Distinct phases of the echo signal can be BS amp E 2005 03 0022 2 2 30 D BioSonics ChapterB Methods B2 amp 3 observed more clearly for hard bottom types The phases of the echo signal re lated to acoustic pulse propagation are projected on the Fig 2 According to Fig 2 the phases of bottom echo for mation can be described as follows e phase I FAZA I attack from the moment the pulse reaches the bottom until the time when the bottom is reached by the back slope of the pulse e phase II FAZA II decay be ginning at the end of attack phase and lasting until the time when the front of the pulse reaches the boundary of the ideal beam pat tern phase III FAZA III release lasting until the time when the pulse completely enters the bottom TID i 4 H T i j X EA Figure B 2 Succeeding phases of the soundin
132. xport the following by highlighting them with your mouse a specific ping a specific sample range of the displayed ping bottom areas only Browse allows you to specify the path for the output file V Note Make sure Stop Analysis After 1 Report in Configuration gt Output Report Filters is dese lected Otherwise you will need to select okay after each report in order to export the entire data file BS amp E 2005 03 0022 2 2 30 31 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Chapterl Getting Started Table 1 4 Import Export File Formats Echogram bmp Bitmap can be imported into most word lll dr d Report CSV Comma separated values can be tabulated Bl eras Map bmp Bitmap can be imported into most word DT4 txt text file Export Data gt Copy Object to Clipboard The command Copy Object to Clipboard found in the EXPORT DATA MENU places an image of the active window into the clipboard The image can then be pasted into another program such as a text editor This function is only available for the Echogram window the report window or the map window y Note if the window has been zoomed into only the zoomed in portion of the window will be copied Zoom out to copy the entire window Export Data gt Export Object to File The Export Object to Clipboard command saves the active window to a user speci
133. y filter is in terms of the per cent of the maximum energy present in the data file and can be described using the following mathematical approximations EP P dx where EP is the energy of the it ping and P is the ping plotted against depth x EP is calculated for all pings in the data file and the highest value of EP called EP is determined The energy of each ping is compared to HE Ppa using the ENERGY FILTER an output report line is created for all EP gt x 11177 where ENERGY FILTER For example the energy filter shown in the figure to the left allows samples with a value of 60 or more EPma to pass and be analyzed by the VBT algorithm The Pings per Report option allows you to specify the step size between pings displayed in the output report window as well any exported report file When the option Report the average of all qualifying pings is selected the algorithm returns the average all pings that pass through the filters in a ping step For example the output for ping 31 when pings per report is 20 is the average of all values over 60 for pings 12 31 When the averaging option is not selected all pings that pass through the filter are displayed The option Stop analysis after 1 report stops the analysis after 1 report If report the average of all qualifying pings is selected the analysis will stop after 1 ping report and the output report window will show 1 report contain
134. y going to Configuration BS amp E 2005 03 0022 2 2 30 10 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT VBT User s Guide 2 Basic Steps gt Bottom Tracking Parameters The Peak Threshold should be a few decibels below the highest peak in the data file Step 2c In the oscilloscope window use the mouse to change the sampling windows vertical lines The sample windows should be adjusted so that the El sampling window the second half of the first bottom echo has red hatch marks in the oscilloscope window See Section 3 1 for more informa tion on manipulating the oscilloscope window y Note Changes made to the sampling windows sampling thresholds or within the method windows AFTER the data has been analyzed played will not take effect until the data file is reanalyzed Step 2d There are 3 ways to select the bottom typing method to analyze the data with i Press the corresponding method button in the toolbar Eme ii Select the the corresponding method from Select Bot tom Typing Method in the CONFIGURATION MENU op 111 Select the the corresponding method from Configure Bottom Typing Algorithms inthe CONFIGURATION MENU We recommend always working with Method B4 Please re fer to Section 3 2 and Section 2 1 2 for in depth information on manipulating the method windows y Note Only one bottom classification method can func tion at a time The data will be classified
135. ying the parameters de Above Bottom Blanking Zone samp ANO Alarm Limit samp Pus hodie scribed in the following ta et i 20logR Tracking Window samp Es ble Please refer to Chapter pe ieri C Aia 4 1 for detailed information regarding the bottom track ing parameters BS amp E 2005 03 0022 2 2 30 25 Copyright C 1998 2005 BioSonics Inc All Rights Reserved PRE RELEASE DRAFT Tip E2 is only necessary when using Method B2 VBT User s Guide 2 Chapterl Getting Started Bottom Tracking Parameters Peak Threshold The minimum intensity an echo must reach for PEAK WIDTH number of samples to be considered a bottom echo Peak Width The minimum number of consecutive samples an echo must be above the PEAK THRESHOLD to be considered a bottom echo Bottom Detection Threshold The maximum intensity a portion of the echo must fall belov ABOVE BOTTOM BLANKING ZONE number of samples to be considered an endpoint of the bottom echo Above Bottom Blanking Zone The minimum number of consec utive samples an echo must be below the BOTTOM DETECTION THRESHOLD to be considered a bottom echo Alarm Limit The number of consecutive lost bottoms before the bottom tracking algorithm is reinitialized and the bottom echo is tracked from the bottom of the last ping s bottom echo Bottom Tracking Window Search window for the bottom tracking algorithm The window is centered on the bottom line of the last found bottom an

VBT 1.10 User's Guide - Ecologia e Gestão Ambiental

Contents

Download Pdf Manuals

Related Search

Related Contents

VBT 1.10 User's Guide - Ecologia e Gest&atilde;o Ambiental

Contents

Download Pdf Manuals

Related Search

Related Contents

VBT 1.10 User's Guide - Ecologia e Gestão Ambiental