We have geo-referenced ortho mosaic images collected for the same area over several years using UAVs, and we're trying to analyze very basic vegetation changes from year to year.
I've been trying to come up with the right search terms to find an area I can study, but am so far unable to do so.
Our images are visible light, so we don't have separate bands.
My hope is to find a way to basically break an image down into "Basically Green" for plant canopy areas, and "Basically NOT Green" for open sand and dirt.
Ideally in the end we would like to be able to draw zones where vegetation used to be and now isn't, and the offset of areas where vegetation now exists where it didn't used to be.
Looking at the images, there are some shades of green that would allow us to do some basic plant classification, but just "alive" and "dead" would be a good start.
Is Supervised Classification the way to go?
Any and all suggestions greatly appreciated.