Issues with Supervised classification in SNAP

Hi there,
I am trying to get a supervised classification to work in SNAP but each time I am getting 100% classification in one class. I have imported an image, resampled and subset it. I have then drawn a number of training polygons on it but when I classify with either RF or maximum likelihood it just gives me one class as 100% and in the frequency table there are just diamonds with ? in.

I read the previous posts about this and the need to reproject which I have done but when I do this I am left with a black screen just showing my training polygons. I reprojected to WGS 84/UTM automatic.

Any help on this would be very much appreciated as I am not sure where I am going wrong here.
cheers
Liz

To go sure, please

  • reproject to Geographic Lat/Lon WGS84 (Sentinel-2 data is already in UTM)
    you can un-check “reproject tie-point grids” to save time
  • add the training data after the reprojection
  • run the classifier again

Hi there,
Thank you so much that seems to have worked. So I brought in the image, resampled to 10m , subset, reprojected to WGS 84 (not the WGS 84/UTM (automatic)) which I had done before, then added the sample data and thats working now.

Thank you so much for your speedy response.

cheers
Liz

very good!