Polarimetric Variability of Simulated Vegetation Ground Cover

Abstract:
It is observed that decompositions of fully polarimetric SAR imagery can display spatial variation across crop fields that are to all intents and purposes homogeneous. Seeking to interpret such observed variation we have constructed a first order scattering model in the distorted Born approximation. The model has been used to perform coherency matrix calculations for a complex vegetation canopy above a rough soil surface. Polarimetric decomposition techniques have been used with simulated coherency matrices, and by varying scene biophysical parameters we have charted trends in decomposition. Polarimetric decompositions display sensitivity to scene parameters such as canopy depth and soil roughness. However the calculations reveal that, consistent with observation, polarimetric decomposition is likely to display greatest sensitivity to crop-type, rather than variation of the studied parameters within crops of a single species.