Statistical assessment of H/A target decomposition theorems in radar polarimetry

Abstract

Target decomposition (TD) theorems have been widely applied in radar polarimetry, serving as the bridge between radar measurements and the physical parameters of a scatterer. The H/A target decomposition is one of the most suitable tools for performing data interpretation in the study and characterization of natural scatters. In this paper, the statistics of the parameters of eigenvalue, eigenvector and H/A have been analyzed via a novel method, and the probability density functions (PDF) have been obtained. It is pointed out that all these parameters can be seen as asymptotically nonbiased with respect to the number of looks. In order to reduce the biases for a small number of looks, a novel estimator for the parameters has been proposed. We give the basic theorems for accurate estimation of parameters.

Model-based target decomposition with the π/4 mode compact polarimetry data

“…One important decomposition approach in the quad-pol SAR is H/α decomposition [12,13], which is based on the eigenvector analysis of T 3 . Here, a key physical parameter is α, which identifies the scattering mechanism in the range 0 α 90 • (α = 0 for the surface scattering, α = 90 • for the double-bounce scattering).…”