Polarization-Based High Resolution Radar Scatterer Classification

We consider the use of polarization diversity for the detection, tracking and classification of scattering centers in high range resolution radar (HRR) data. Specifically, we propose a novel exploitation of polarization ratios to label moving scattering features into various geometrical classes including edges, tips or smooth reflectors. We demonstrate scattering-center labeling for both parametric and non-parametric classification algorithms. Experimental results on simulated HRR data are provided and indicate that polarimetric information can be exploited for geometrical typing of scatterers