I am trying to assess the accuracy of multiple classifications using the "Create Accuracy Assessment Points" tool in ArcGIS Pro 2.4.2. I classified each image using 3 different classification algorithms into "leaves", "soil", and "shadows." I have plots like the ones pictured below, to which 500 points were randomly assigned. I then went through and labeled the feature that each point fell on (without looking at the classified image) and labeled the points as either "leaves", "shadows", or "soil."
I should say that the "Create Accuracy Assessment Points" tool picks one of the classified rasters and puts points on the raster, which I then hid from view and used the original RGB image to label my ground-reference points. Therefore, after ground-referencing, you will have two columns in the attributes table: The classified points and the ground-reference points. You can then compute an error matrix to assess the accuracy of your classification. Here's classified raster 1, so that I can more clearly explain what is happening:
Let's say I want to use these same points to label the other classified rasters, so that I don't have to re-label them for a different image. I want my ground-reference labels to carry over to the next raster, while generating a new set of labels for where the points fall on the other rasters. How can I do this?