Analysis of PolInSAR precision for forest and ground parameters estimation in tropical context

The analysis presented in this paper is focused on the TropiSAR PolInSAR, acquired over a dense tropical forest area. Under canopy Ground height estimation based on PolInSAR analysis is explored and shown to be in good agreement with the LIDAR data. The simple estimation based on HH phase center is observed to overestimate the ground height by about 10m and presents a larger standard deviation than the Polin-SAR approach. The single baseline approach for vegetation height estimation performs well when the altitude of ambiguity is favourable with a RMS error ranging from 3 to 5 m, depending on the topography. In order to analyse the effect of topography, we have investigated the relative contribution of the ground and volume layers, obtained based on the Tabb maximum likelihood approach. As expected the topography is mostly affecting the ground contribution, both in term of backscatter variation and of scattering mechanisms. These in turn will influence the estimation performances as given by the Cramer Rao Lower Bound analysis.