Rule Based System for Thematic Classification in Synthetic Aperture Radar Imagery
John C. Curlander, Woody Kober · 2005
The structure and initial results of a symbolic, model-based approach to feature classification in Synthetic Aperture Radar (SAR) imagery is pre- sented. The. models include quantitative, qualitative and relational as- pects of feature signatures. The prototype system consists of a combina- tion of a 120 rule knowledge-based analysis tools, and a 20 procedure data extraction library. The initial testing over the four SAR data set re- sulted in a 71% correct classification rate in a fully automated mode.