POLARIMETRIC OBSERVATION OF FOREST COVER IN RADAR REMOTE-SENSING

In this paper, and from the received waves, we investigate a number of polarimetric radar techniques to characterize the polarization respon se of forest cover, using the example of the Landes Forest. Magnitude, as well as phase difference and amplitude ratios between the polarime tric channels, which are closely related to the texture of the scene, offer reliable means for data characterization and discrimination purp oses. The statistical properties of high resolution polarimetric SAR a re investigated, using the K distribution to model the received amplit ude of each area selected. This distribution is derived with a paramet er, alpha, which is closely related to the heterogeneity of the ground cover, and thus provides a good tool for discriminating among differe nt regions. From the Stokes matrix, the co- and crosspolarized signatu res are constructed, for both clear cut and full-grown tree regions, i n order to identify the predominant scattering mechanisms contributing to total measured backscatter.