I am trying to use the Spectral Unmixing feature in SNAP on a S3 subset. I have created a tab separated csv file using python in several different ways. The csv endmember file uploads successfully and the individual endmembers can be seen in the LSU popup window. However, when I click Run, I get error saying “Matrix is Singular”. I am not sure what this means or how to address the problem.
If I create a tab separated csv file with the same data using Excel the LSU tool runs successfully. I am not sure what difference using python is creating.
I am also having problems with the Endmember file for the unmixing. I exported my endmembers from Envi in ASCII format and then converted it in a tab separated CSV file using Excel and Notepad. It looks like this:
There is only a slider missing in the list of endmembers. If you further step down through the list by the cursor keys you see different spectra highlighted in the graph.
Also, it would be good if you have a different name for each of the endmember.
And you have to make sure that you have more source bands then endmembers.