Principles and Applications of Polarimetric SAR Tomography for the Characterization of Complex Environments
Résumé
Despite its widely recognized capabilities for mapping and characterizing large areas, 2-D Synthetic Aperture Radar (SAR) imaging meets serious limitations over volumetric media, due to its incapacity to discriminate scattering contributions in the elevation direction. This paper proposes some methods for characterizing complex volumetric environments using polarimetric SAR tomography, a 3-D imaging technique based on the use of diversely polarized electromagnetic waves acquired from different trajectories. The use of polarimetric diversity permits to both improve the tomographic separation between different components of complex volumetric media and to characterize the EM behavior of the observed environments. A set of spectral estimation techniques, adapted to tomographic focusing, are tested against signal models accounting for the statistical complexity of hybrid volumetric environments. Due to their statistical adaptivity, their robustness to mismodeling and their accuracy, spectral estimators based on weighted subspace fitting criteria are selected and extended to the polarimetric case. The effectiveness of the proposed approaches is assessed over real data and for three different applications, related to urban area 3-D mapping using a minimal set of images, tropical forest structure characterization using low frequency waves, and under-foliage concealed vehicle imaging.