These are examples of image classification in ENVI. On the left is ENVI’s automated (“unsupervised”) classification and on the right is a manual (“supervised”) classification. ENVI’s automated classification is very good. It’s quick and easy to get good results – these results could be even better had I fine-tuned the settings a bit, but in this case I just ran the default settings (with the exception of iterations – the default is 1, I ran 6). The whole process took about 2 minutes. The image on the right is the result of manually selected classifications. This approach allows a lot more control and given time, a skilled analyst could generate very accurate classifications. These results are after about 10-15 minutes of selections and include a few areas that are better classified than in the automated process, for example the bright red spot at about 300° below and to the right of center is a burn-area that the automated process did not pick up.