Quantitative Analysis of The General Polarimetric Model-based Decomposition Technique by Using Monte Carlo Simulations

In this paper, we quantitatively analyze the feasibility and robustness of parameter inversion based on the general polarimetric model-based decomposition method proposed recently by Chen. To improve the performance, we propose some modifications on the optimization algorithm, including redefined boundary conditions, transformation of variables and parameters initialization. A number of Monte Carlo simulation tests are performed to analyze the feasibility of the proposed improvements. Results show that estimation accuracy is significantly increased. In addition, this study also points towards the necessity of checking the quantitative accuracy of model-based PolSAR techniques beyond their proven utility for qualitative features extraction.