摘要：

This dissertation concerns different topics with the aim to improve the understanding of polarimetric SAR interferometry (PolInSAR) and to provide techniques for vegetation parameter retrieval for remote sensing applications. Theoretical studies are conducted to rigorously define and analyze the PolInSAR coherence set, and to develop and interpret multi-baseline coherence optimization techniques. A phenomenological polarimetric interferometric model, designed for geophysical parameter retrieval, is derived for volumetric media over ground. For forest vegetation observed at L-band, this model accounts for the ground topography, canopy layer and total tree heights, wave attenuation and refractivity in the canopy, tree morphology in the form of the orientation distribution and effective shapes of the branches, surface scattering contribution, and double-bounce ground-trunk interactions. A parameter retrieval framework is developed for single- and repeat-pass acquisitions which is, in the latter case, robust against the temporal decorrelation and permits its estimation in every baseline. The parameter estimation performance is evaluated on simulated and real airborne L-band SAR data in both single- and multi-baseline configurations.

展开