User`s manual
Contents
1. pointed upwards Most smartphones have a built in camera both on front and rear side of the phone It does not matter which camera will be selected It is then possible to rotate or flip the taken photograph Fig 9D On photographs taken with a narrow angle lens the definition of the hemispherical border has not relevant Fig 9E Such a function is important however when a fisheye lens is used Fig 4 and 5 The hemisphere centre and diameter are usually defined using the first photograph Fig 9F Also lens projection is important for correct estimation of the Canopy Cover Index Fig 9G Lenses of built in cameras usually have a polar projection the shortest distance between two different angles is the same in each part of the photograph while external fisheye lenses can have angular projection the shortest distance between two different angles is smaller at higher zenith angles A defined artificial mask in our example 69 helps to make the photograph independent from camera orientation Fig 9H because the photograph uneven width and height is restricted by a circle with the same distance from its centre This page also supports a selection of the level of precision how many pixels from the photograph will be analysed and colour channels red and blue colours do not penetrate through the vegetation Final definition of light gaps Fig 91 depends on the proper selection of a cut level between white and black pixels The2. a hemispherical photograph as the weighted sum of black pixels as fractions of all pixels within the photograph segements The weight of these area fractions is a function of the zenith angle and depends on the camera lens projection lens geometric distortion used Canopy Openness Measure simply complementary to Canopy Closure Canopy Cover Index the main result of hemispherical photograph analysis within this application Canopy Cover is the perpendicular projection of tree crowns onto a horizontal surface Estimation of canopy cover by visual observation is commonly used in vegetation science and forest ecology for description of vegetation structure The Canopy Cover Index can rescale hemispherical projections of photographs taken from a single position to perpendicular projections of light gaps Therefore the index is suitable to estimate canopy cover directly from hemispherical photographs It was found that the index need not be calculated from the fully hemispherical photograph taken by fisheye lenses If only the forest vegetation above the plot is analysed i e horizon mask of 45 70 is defined the results are more precise Selected Colour Cut Level ASCII value of the colour in a grey scale as the boundary between light and dark pixels 0 255 10 Recommended sampling design Estimation of canopy cover from hemispherical photographs depends on both employing a satisfactory field sampling strategy to capture representativ
3. final result is presented in the Results page Fig 9J 3 BL Ao 8 18 30 E T MLA 18 30 IGLAMA Gap Light Analysis Mobile App GLAMA Gap Light Analysis Mobile App Rot tate ard transpose image Fisheye crcie visibility Define cur leve T BLAA mi 12 31 T DLA vo 8 18 31 aN 19 3 ah d 30 18 3 DL ad ws 8 18 33 Gap Light Analysis Mobile App b h FILE GLAMA_20150 27_18262 Fig 9 Sequence of steps during field image analysis Note Moving between pages is also possible in both directions by moving a finger through the display to the left or right 8 Defining cut level between black and white pixels The value of this cut level significantly influences the result Therefore the application has several tools for proper identification of shadows made by vegetation and the incoming light from vegetation gaps The application calculates a histogram of grey colour frequency the real colour mix depends on selected channels Fig 10A The cut level is defined from the range 0 255 and the value automatically recommended corresponds to the intersection of the ideal normal distribution of black and the normal distribution of white pixels Fig 11 The cut level can be modified manually by moving the horizontal bar at bottom of the layout Fig 10B A single tap to the image will zoom the area touched by the finger Fig 10C Inverse masking of vegetation by red colour may identify whether t
4. out of the page view and no mask 4 hemispherical photograph with edges partly out of the layout view and with 45 artificial horizon mas 5 partly hemispherical photograph taken by built in camera with edges fully out of the page view and 6 partly hemispherical photograph taken by built in camera with edges fully out of the page view and with 69 artificial horizon mask
5. 9 91 Selected Colour Cut Level Canopy Cover Index 60 01 Gl FILE GLAMA_20150627_182629 Results No of Light Dark Pixels 74 Part of Hemisph Taken by Camera 61 61 Canopy Openness Canopy Closure 7 15 Selected Colour Cut Level Canopy Cover Index 85 80 234 RAL GUAMA Gap Light Analysis Mobile App Sky area definition Define cut level Light RAL CLAMA Gap Light Analysis Mobile App Sky area definition Define cut level GLAMA Gap Light Analysis Mobile App Sky area definition Define cut level Dark oye FILE 30341 C DSCN1488 Upraveno Results Part of Hemisph Taken by Camera 1 To Canopy Openness Canopy Closure 1 63 92 32 Selected Colour Cut Level Canopy Cover Index 67 33 205 RAL FILE GLAMA_20140312_104446 Results Part of Hemisph Taken by Camera T Canopy Openness Canopy Closure 6 30 93 10 Canopy Cover Index Selected Colour Cut Level 66 37 Sey Ball 100 13 28 FILE GLAMA_20150627_182629 Results No of Light Gap Fraction of Selected Area Part of Hemisph Taken by Camera 61 6 Canopy Openness Canopy Closure 4 26 95 74 Canopy Cover Index Selected Colour Cut Level 87 57 234 Fig 12 shows six different situations 1 fully hemispherical photograph with no mask 2 fully hemispherical photograph with 45 artificial horizon mask 3 hemispherical photograph with edges partly
6. GLAMA Gap Light Analysis Mobile App User s manual application version 3 0 Lubomir Tichy Dept of Botany and Zoology Masaryk University Brno Czech Republic 1 Introduction Gap Light Analysis Mobile App GLAMA is a program for calculation of the Canopy Cover Index estimating canopy cover from hemispherical photographs The program which is freely accessible from the Google Play website can be used for hemispherical wide angle and standard photographs with known lens angle of view The program was primarily designed for use in the field but can also analyse hemispherical photographs saved on external storage Select picture source Sky area definition Define cut level et Gee 95 45 Fig 1 Simple sequence of steps to calculate Canopy Cover Index 2 Definition Canopy Cover is the perpendicular projection of tree crowns onto a horizontal surface Estimation of canopy cover by visual observation is commonly used in vegetation science and forest ecology for description of vegetation structure 3 Efficiency The Canopy Cover Index is a quick and robust method for precise canopy cover estimation comparable to visual canopy cover estimation but unaffected by observer bias Not only can it be used on already captured photographs but the index can also be employed on smartphones by using the GLAMA Android application to rapidly capture hemispherical photographs and immediately calculate their index values d
7. and proper definition of lens projection at the next screen prepare the application for correct calculation of the Canopy Cover Index D Sky area definition Define cut level TRAN A CIAMA Cap Light Anal MA Gap Light Analysis Mobile App Horizon definition preciseness channels 8 16 Next Sky area definition Define cut level L Horizon definition preciseness channels 1 4 Preciseness analysis speed Preciseness analysis speed All pixels 1 4 All pixels 1 4 olour channels olour channels Blue Bluetre Blue Blu Fig 8 Second option for definition of hemisphere diameter and two alternatives for definition of horizon masking Whereas the first two images A B show a simple mask defined by one angle above the horizon in all directions the last two images C D show the possibility of individual masking in 16 different directions The light incoming from higher zenith angles travels through scattered vegetation decreasing the precision of canopy cover estimation In addition when conducting a visual assessment of canopy cover the observer s view of the canopy comprises only an upper portion of the sky hemisphere Artificial hemisphere masking helps to remove the outer area of a hemispherical photograph eliminating some of the light from higher zenith angles Tab 1 shows horizon angles which can be used for delimitation of a hemispherical photograph depending on the vegetation height and sa
8. e images and on proper pixel classification of the photographs The field sampling strategy must avoid potential bias in canopy characterization that can occur due to patchiness of the forest canopy especially if images are captured only from openings between trunks This bias can be avoided by using approaches such as collecting images from multiple points in transects or grids and averaging their data Proper pixel classification is also essential to the accurate use of hemispherical photograph analysis to quantify canopy cover The GLAMA software provides a cut off level between black and white pixels based on histogram parameterisation and manual visual evaluation of pixel colour differences Therefore you should check every time to make sure that the cut level is selected properly i e light gaps are not identified as vegetation nor vice versa mm A GLAMA Gap Light Analysis Mobile App Sky area definition Define cut level Lisht Lig t AAL CLAVA Gap Light Analysis Mobile App Sky area definition Define cut level Uklac C dis e GLAMA Gap Light Analysis Mobile App Sky area definition Define cut level RAS FILE ID341 C_DSCN1433_upraveno Results Part of Hemisph Taken by Camera 10 Canopy Openness Canopy Closure 9 34 90 66 Selected Colour Cut Level Canopy Cover Index 141 205 TRAG FILE GLAMA_20140312_104446 Results 15 5 03 Canopy Openness Canopy Closure 10 09 8
9. h built in smartphone camera To my knowledge it is not possible to take a truly hemispherical photograph Fig 2A with any smartphone camera at present Such a photograph can be taken however with a professional camera with an expensive fisheye lens The analysis of a photograph from such a camera is also available but the photograph should first be downloaded to the storage card of the smartphone see chapter 6 Fig 2B represents a photograph taken with a smartphone with additional fisheye lens Fig 3 The analysis of part of the hemisphere can be done even with a built in lens such as found in any smartphone Fig 3 External inexpensive fisheye lens for smartphone Lens calibration for photographs in which the hemispherical border is visible Fig 2A and 2B is quite easy and does not require a special procedure After the image is taken the border between a real shadow and the lens masking can be identified by definition of the grey colour cut level Fig 4 B C RAe or gt Ale ro GLAMA Gap Light Analysis Mobile App GLAMA Gap Light Analysis Mobile App Fisheye circle visibility Define cut level Fisheye circle visibility Define cut level Pee eee es ee ee ee ee t RAAL GUAMA Gap Light Analysis Mobile App Fisheye circle visibility Define cut level Again Refresh Fig 4 Better visualisation of hemisphere border A horizontal bar below defines a cut level of grey colour with a range from 0 to 255 Whe
10. he cut level was selected correctly El B alal 30 18 32 A Babul 39 18 32 fa El F Bi Lal af se 18 33 A PAET 18 33 GLAMA Gap Light Analysis Mobile App GEAMA Gap Light Analysis Mobile App Sania C3 se Dect yao Fe GLAMA Gap Light Analysis Mobile App Sky area definition Define cut level Sky area definition Define cut level Sky area definition Define cut level Sky area definition Define cut level ata ra A os von reresn no a oi Es e i ES EI 2 Fig 10 Tools for definition of cut level between black and white pixels Note Every time you change the cut level use the Refresh button to recalculate it Otherwise the results will be calculated from the last refreshed state o Oo minimum equal frequency proportion o of pixels Oo a st _ Oo N J black pixels white pixels Se T T T T T T 0 50 100 150 200 250 0 255 Fig 11 Simulated distribution of pixel colour frequency in grey scale 0 255 from analysed image The image has to be divided into black and white pixels which are log normally distributed over the gradient The definition of the correct cut level is the main task influencing estimation precision We can estimate the frequency of black or white pixels as their theoretical log normal distribution around some mean value If the shape of both distributions differs the modelled equal pr
11. irectly in the field 4 How to install the application 4 1 Android Operating System Open the official Google Play web address https play google com store apps details id com mobilesglama and install the program automatically to your device Alternatively the APK file can be downloaded to your smartphone or tablet from http www sci muni cz botany glama You should run the APK file manually after the download 4 2 WINDOWS Operating system If using a PC operating in Windows this Android application should be installed to any Android emulator Of these I have had experience using BlueStacks which supports running the program directly on a PC for a posteriori analysis of photographs captured by a camera in the field 5 Lens calibration definition of hemisphere centre and edges Proper calculation of the Canopy Cover Index requires the definition of the hemisphere perimeter red circle in following photographs in pixels There are three examples of photographs Fig 2 in which different parts of the hemisphere can be taken by the camera ight Analysis Mobile App Lens projection Select type t Lens projection Select type Fig 2 Different types of photographs which can be analysed by GLAMA Left photograph was taken with external Nikon Coolpix camera with true fisheye lens and 180 view Centre photograph was taken with external fisheye lens used with smartphone built in camera Right photograph was simply taken wit
12. le hemisphere See lens calibration section The application can also be used as a tool to train students and plant ecology researchers inaccurate estimation of forest canopy cover in the flald Fig 6 Calibration of narrow angle usually built in lens of a smartphone camera It is possible to use any circle such as a lampshade in your office After manual definition of the circle centre and diameter the program then automatically calculates the circle s diameter in pixels it is possible to calculate a virtual diameter of the whole hemisphere Remember or record hemisphere diameter and recommended angle of horizon mask for their next use in the calculation process The hemisphere diameter can be defined using either of two display layouts 1 In the page with definition of hemisphere borders Fig 5 the central X button must be pressed The circle will be centred and a query will ask for the circle diameter Fig 7 or 2 the second option is in the page with the definition of the horizon mask Fig 8A where hemisphere diameter can also be edited B kd Lu Yo F Badal 24 8 17 24 GLAMA Gap Light Analysis Mobile App Define borders Write circle diameter in pixels 1014 a Fig 7 The sequence of steps by which to centre a circle define the virtual hemisphere and define its diameter Whereas the circle size and position are wrong in the first image A after centring the proper diameter definition
13. mpled plot size Maximum height of canopy cover 15m 20m 30m 40m 10 x10m 65 71 77 80 Sample plot size 15x15m 55 63 71 76 20 x 20m 47 55 65 71 25 x 25m 41 49 60 67 Tab 1 Approximate estimated angles of horizon masking for different plot sizes and different heights of forest canopy when the hemispherical photograph is taken from the plot centre 6 Paths used by the application The main path of the application is internal storage images All photographs taken by a built in camera in JPG format are stored to this path Their names start with GLAMA and continue with the date and time of the capture of the photograph Each photograph is followed by an ini file of the same name which is created during the analysis process It saves all selected parameters and repeated analysis of the photographs can use these pre defined values Note All hemispherical photographs taken by a camera not in a smartphone must be copied to this path of internal storage before the analysis The subfolder internal storage images note contains a global ini file which helps with application of all parameters from the last image analysis to a newly taken hemispherical photograph 7 Example of field analysis The process of image analysis in the field starts after selection of the From camera button Fig 9A the photograph is then taken Fig 9B and 9C The camera must be held in the horizontal position
14. n you move it and select the Refresh button the dark part of the image will be covered by red dots for better identification and definition of hemisphere borders see Fig 5 Next the proper position and diameter of virtual border circle must be defined Position and diameter size are changed simply by touching the display and moving a finger over it Fig 5 RAL n TRAE TRA 4 u GUAMA Gap Light Analysis Mobile App 7 77 CLAMA Gap Light Analysis Mobile App GLAMA Gap Light Analysis Mobile Ap Define borders Define borders Define borders Fig 5 The page with definition of hemisphere borders can be done with a finger Select the option Center and drag the finger to the centre of the photograph Then select the option Radius and drag the finger to the edge of the visible part of the photograph If the lens view is so narrow that borders are not visible Fig 2C it is necessary to apply the second method 1 Tap on the Question mark button in the Main display Fig 6A Select the Lens calibration button Fig 6B 3 Take a photograph of any circle of known diameter from a known distance It does not matter if it is a circle of known size printed on paper or some such circle as a cup plate lampshade etc Note Record the circle exactly in the photograph s centre Define the border and the centre of the photographed circle Fig 6C 5 Write Fig 6D the circle diameter in cm distance of
15. oportion of black and white pixels green line need not be equal with the minimum frequency over the colour gradient The application proposes the cut level between black and white pixels as an intersectin of both distributions as the value equal proportion of pixels 9 Interpretation of results Explanation of outputs Fig 11 Total No of Pixels in Circle Frame The circle defines the virtual border of the hemisphere hemisphere horizon in the photograph The application estimates the number of pixels within the whole circle or hemisphere In cases when only part of the whole hemispherical circle is visible on the page the number of visible pixels is presented in brackets Note The number of pixels in the circle frame is counted independently from the defined mask which does not influence the results Number of White Black Pixels The real number of white black pixels in visible frame Note The number of white black pixels is limited by the defined mask All masked white pixels are counted as black Gap Fraction of Selected Area Ratio between white and all pixels counted within the visible part of the circle Part of Hemisph Taken by Camera It is the percentage ratio of the visible part of the hemisphere Canopy Closure The proportion of the sky hemisphere obscured by canopy elements when viewed from a single point Canopy closure is calculated from
16. the circle from the lens in cm and the lens projection planar or angular Recalculate the hemisphere diameter pixels and recommended angle of the horizon mask A B D vB Abs roe m Boe TIE GLAMA GLAMA i ap e GLAM Gap Light Gap Light Analysts Mobile App Gap Light Analysts Mobile App Rralysis MoblloApp Select picture source l i i i ct picture Sourc Calibrate lens Calculate full hemisphere diameter a oe Gap Ligh Mysis Mobile App GLAMA is a free progtanifor analysing hemispherical wide angle alld standard photographs with 199 known angle W lens view The photographs api ta than E E smartphone taet camera or as JPEG files on a computer with Android operation system or Android simulafer 7 This means that the application can be used either directly in field research or to analyse i stored hemispherical photographs previously A MT taken with a digital camera Most parameters need to be set only once with the subsequent photographs automatically subject to most of the previously defined parameters Therefore tho analysis of a photagraph takes only a few ph diamejer 1007 pixels seconds If photographs are taken without a fisheye lens 69 1 itis necessary to first measure the real view S angle of the camera to assess level of artificial masking of the portion of the hemisphere that is not photographed and recalculate it in pixels of the photograph to the radius of the who
Download Pdf Manuals
Related Search