We have many orthographic images that we are using ENVI 5.2 to do supervised classification. They all share many of the same characteristics and are form the same region. We want to do some supervised learning on a few images, but carry over the results and then automate the process. We have far too many images to do them all by hand. We would like to train on a few, automate the process, check some by hand, and add new classification data as needed to resolve mistakes. Any idea how to do this? If this is not feasible, then the software is not very useful.