Polarimetric SAR Tomography for forest structural estimation

This paper investigates the use of Polarimetric and Tomographic Synthetic Aperture Radar (PolTomSAR) data for forest vertical structure retrieval. In order to obtain robust estimation, a Multi-Baseline Random-Volume-over-Ground (MB-RVoG) modeling of the forest is used. This model allows polarimetric and spatial separations of ground and volume contributions and then makes easier the estimation of their geophysical parameters. However, ground effect may be a major limitation factor on vegetation parameters estimation. At P-band, the measured signal does not depend only on the canopy layer but also on the vegetation-ground interactions. To overcome this problem, a generalization of RVoG model is introduced in this paper. This generalization consists on taking into account both of ground canopy and ground on vegetation response. This new modeling leas to an accurate retrieval of forest parameters and then accurate estimation of vertical structure. Afterward, the model is validated on P-Band PolTomSAR data acquired during the TropiSAR campaign over the test site of Paracou in French Guiana.