Parameter Estimation for the K-distribution in PolSAR Imagery Based on Hybrid Moments

The K-distribution is usually used to model the Polarimetric Synthetic Aperture Radar (PolSAR) imagery. The parameter estimation method for K-distribution is very important,which affects the fitting degree of the model. However, the classical method of matrix log-cumulants relies upon a nontrivial inversion of a nonlinear equation, which introduces a computationally expensive stage into the estimation procedure. Moreover, the bias is large when the sharp parameter is smaller than 1. Therefore, a new method for estimating the sharp parameter of K-distribution based on|z|rlg|z| is proposed. This method is more adaptable to parameter estimation under different sharp parameters, and the performance is better than matrix log-cumulantes whenis small. In addition, the proposed estimator has an analytical expression at r=1/d, which allows rapid caculation. Finally, the estimation accuracy of this new estimator is compared with previous ones through simulation data and real data. The results show that the new estimator is effective and robust, which is of practical value in solving the accurate parameter estimation of K-distribution.