I have been trying to come to terms with the supervised classification tools within SNAP for some time. I am trying to correctly classify crops and ultimately want to be able to classify and then identify a specific crop across any image in Sentinel-2 but I cannot seem to find a way to do this.
To date I have managed to isolate the broad spectrum of the plant I wish to identify by using EO imagery as well as NDVI in conjunction with known GPS coordinates of a crop. My understanding is that in SNAP I can create data bins specifying which areas I know are the crop in question and which aren’t (urban, trees, water etc) and then run a supervised classification on a Sentinel-2 tile. However, can I use this process to then look at another geographically separate location to identify that specific crop without having to go through the process of training again?
I essentially want to be able to identify a specific spectral signature within any Sentinel-2 tile. Is this possible?