Free download of the recent review 'An overview of Digital canopy photography in Forestry' Is available at (limited to 50 downloads): https://www.nrcresearchpress.com/eprint/HRQQHYBTIS9VPMEQG9QI/full
If you have hundreds of photos, than the WinSCanopy may be the choice (it performs batch analysis). Otherwise, freeware GLA provides comparable functionality.
In my experience Winscanopy contains the most comprehensive set of tools to process hemispherical photos, and it has batch processing capabilities. Currently the major issue is the software is not freely released. More cheap alternative is hemisfer with a student licence, or the freeware CANEYE. GLA has manual thresholding, which is not the best way to obtain replicable results.
As an alternative to hemispherical photography, I would suggest you to consider cover photographic method developed by Macfarlane et al. 2007 Agric For Meteorol 143. It is simple, fast and reliable, and relatively insensitive to skythe condition and camera exposure. Take a look at the enclosed papers
I have compared estimates from GLA, CANEYE and HEMIVIEW with actual pine LAIs calculated from biomass assays, and CANEYE was allowed me to produce the most precise calibration model.
Hi, as addendum of my previous comment, I enclosed an article where yo can find URL to download a freeware (Canopy Cover CaCo) to process cover photography images -
Hi, you could als check the App GLAMA - Gap Light Analysis Mobile that is running on your mobile device. It can be an advantage to instantly gain results in the field.
I have recently played with the 'caiman' package of R (https://github.com/GastonMauroDiaz/caiman) and it is very good. I found it correctly classified fisheye images, even when canopy and sky pixels are not well-contrasted.
For non-hemispherical images, particularly restricted FOV images such as cover photography (Article Estimation of Leaf Area Index in eucalypt forest using digit...
), you can use more general purpose packages to binarize canopy images and then you can directly invert classified gap fraction to get canopy attributes such as canopy cover or LAI (see the above cited paper).
I suggest to take a look at the "authothresholdr" package - it contains sixteen thresholding methods translated from another Auto threshold plugin of ImageJ. Some of these thresholding methods have been already tested in canopy images (see Article Standardizing the Protocol for Hemispherical Photographs: Ac...
;
Article An objective image analysis method for estimation of canopy ...
).
Also the recent bwimage seems interesting: https://f1000research.com/articles/8-1168/v2
Among all the free version of hemispherical photographs analysis software, CANEYE is the best... It is very interactive and very effective to discriminate the sky and vegetation.
Following some suggestions, my approach was to first correct photographs for geometric distortion using the correspondent camera and lenses profile available in Rawtherapee v5.7, a freely available image editor. Then, in R, I have separated the RGB channels and calculated the greenness index (Chianutti et Al., 2014). Images were then binarized (0.5 threshold) and I calculated the percentage of white pixels for each image, as a proxy for canopy openness.
Free download of the recent review 'An overview of Digital canopy photography in Forestry' Is available at (limited to 50 downloads): https://www.nrcresearchpress.com/eprint/HRQQHYBTIS9VPMEQG9QI/full