Adaptive Model-Based Decomposition of Polarimetric SAR Covariance Matrices

Abstract

Previous model-based decomposition techniques are applicable to a limited range of vegetation types because of their specific assumptions about the volume scattering component. Furthermore, most of these techniques use the same model, or just a few models, to characterize the volume scattering component in the decomposition for all pixels in an image. In this paper, we extend the model-based decomposition idea by creating an adaptive model-based decomposition technique, allowing us to estimate both the mean orientation angle and a degree of randomness for the canopy scattering for each pixel in an image. No scattering reflection symmetry assumption is required to determine the volume contribution. We examined the usefulness of the proposed decomposition technique by decomposing the covariance matrix using the National Aeronautics and Space Administration/Jet Propulsion Laboratory Airborne Synthetic Aperture Radar data at the C-, L-, and P-bands. The randomness and mean orientation angle maps generated using our adaptive decomposition significantly improve the physical interpretation of the scattering observed at the three different frequencies.