Polarimetric SAR tomography of natural environments using hybrid spectral estimators

SAR tomography is the extension of conventional two dimensional SAR imaging principle to three dimensions. In order to improve the vertical resolution with respect to classical Fourier-based methods, high resolution approaches are used in this paper to perform SAR tomography. Both nonparametric spectral estimators, like beamforming and Capon and parametric ones, like MUSIC, maximum likelihood, are applied to real data sets and compared in terms of scatterer location accuracy and resolution. This paper addresses the discrimination of coherent scatterers presented in the natural environment and a joint approach of estimation and detection is proposed.