A simulation study of topographic effects on polsar classification of forests and crops

We investigate the dependence of polarimetric decomposition techniques upon terrain and vegetation canopy properties. We establish a physics-based model which highlights the interdependence of scattering mechanisms, and employ a simulation which encapsulates this combination of scattering theory and Synthetic Aperture Radar (SAR) imaging theory. We use the simulation to calculate polarimetric SAR images of forested terrain with high fidelity, and perform polarimetric decomposition analysis on the simulated data. Correcting for terrain azimuth slope leaves a residual dependence in polarimetric response that can be interpreted through the dependence of scattering on the terrain and imaging geometry. We will apply a similar analysis to POLSAR data from RADARSAT-2 and PALSAR to collaborate the model predictions.