I have an imagery for study area, and I need to apply segmentation tool in GIS and object-oriented image analysis in “imagine” Remote Sensing.
It must create training sample and signature file to use in Train classifiers, with different multi-resolutions to help assign features in the right pixels. The classes in training sample are;
Paved Roads, Unpaved Roads, Bigger Houses, Smaller Houses, Factory, Business, Construction Sites, Parks, Trees, Agriculture, Desert, Water and Swamp. The classifiers are; Maximum Likelihood, Random Tree, and Support Vector Machine.