Feature Extraction of Ground Vehicles from High Resolution SAR Imagery Based on Electromagnetic Scattering Characteristic

Feature extraction is an important phase in SAR imagery interpretation and target recognition, electromagnetic scattering characteristic is the inherent characteristic of target on SAR imagery. High resolution SAR imagery interpretation and target recognition, particularly the ground vehicles recognition, needs to be carried through based on target scattering characteristic and target's electromagnetic feature extraction. This paper analyzes the electromagnetic scattering mechanism of typical ground vehicles firstly, and the feasibility of modeling the electromagnetic scattering characteristic of ground vehicles with attributed scattering center model has been discussed in brief. And then, by making some improvement on the existing method of attributed scattering center feature extraction from SAR imagery, a modified method for feature extraction of ground vehicles from measured high resolution SAR imagery based on electromagnetic scattering characteristic has been proposed. Finally, performance of the proposed algorithm has been validated on measured 0.3m resolution MSTAR and 0.1m resolution MiniSAR SAR imagery data.