I want to realize a project with spectrally unmixed Landsat imagery. However, I find no satisfying information on the respective GEE function, image.unmix().
I processed a few scenes with the exact same three endmembers both in GEE and with the help of RSToolbox package's mesma() function in R. The results differ largely, with mesma() looking quite more promising. I cannot compare to the widely used ENVI function right now, but it seems that there is strong differences in implementations.
I am not even sure if they all three could classify as Multiple Endmember Spectral Mixture Analysis (MESMA) or if I'm comparing apples with oranges here.
The GEE help file linked at the top quotes:
For more flexible methods, see the Array Transformations page
I fear, I don't quite understand what Array Transformations are capable of and how they are used. How would I build my own unmix() with these and would it enhance my results?
Can anybody could shed light on potential differences and the exact functionality of the GEE implementation?