Scene classification of SAR images acquired from antiparallel tracks using evidential and rule-based fusion

摘要

This paper presents a method for fusion of information derived from different airborne synthetic aperture radar measurement processes and from different observations of the same scene. Dempster-Shafer theory will be used to fuse radar backscatter and phase measurements. Assigned evidence and conflict measurements will be further fused in a rule-based fusion system. the rule-based fusion system combines the Information acquired by an airborne radar instrument imaging the scene from different flight tracks. Results of the fusion system are Presented for a test site in Bavaria, Germany, where remarkable improvement compared to a single observation was observed.