This lab step through using ERDAS IMAGINE to classify imagery using unsupervised and supervised classification.
To start the lab provided a high-resolution aerial image of University of West Florida, which was classified into 50 classes. Then it was necessary to accurately label the classes and group them into five categories. Once labeled a Reorder tool was used to save the images with only those five classes. See snippet below,
The lab then step through the process of supervised classification by annotating an image's spectral signatures. The ERDAS IMAGINE used this signature file to process and classify the images. Optionally, you can create a distance image, to visual evaluate how well the imaged had been processed. See the map below.