Performance Analysis of Multilook K-Model Parameter Estimators Using the Approach of Approximate Expansion

Parameter estimation is a crucial task for accurately model the synthetic aperture radar (SAR) clutter with various terrain classes. K-distribution has been proved to be a good model for describing the statistical properties of clutter in multi-look synthetic aperture radar (SAR) images. Methods of estimating the parameters of multilook K-distribution are briefly reviewed and nine different moment-based estimators are extended to modeling the multilook SAR imagery clutter. Moreover, analytic s of the biases and variances of estimators are derived through the approach of approximate expansion. Based on the performance comparison and analysis, some basic guide lines for estimator selection under different circumstances are also given, and the results show that the performance of the estimator depends on the number, number of samples, the scope of order parameter and the estimator of the amount of calculation. In practice, we should on the basis of the specific situation to choose suitable estimator.