The approach taken is to estimate a set of target features which describes the target as set of attributed scattering centers. Each scattering center is characterized by its relative range, amplitude, and a polarization ellipse which represents the physical structure of the observed target. First, we represent the fully polarized radar backscatter signal measurements as a set of down range scattering centers parameterized by a modified Prony modeling technique. Next, we use the relative range, site, and the resulting parametric scattering centers to form a description of the "structure" of the target. Finally, we use these "structural description" of the measured backscatter signal to classify the target by hidden Markov models (HMM).
展开▼
