A survey of density estimation for sar images

Modeling of synthetic aperture radar (SAR) images has been an important topic of research since the inception of SAR satellites. Many theoretical and empirical models have been presented in literature to accurately model the amplitude SAR images. The method of parameters estimation of the probability density function (PDF) for selected models is another topic of research associated with modeling. Earlier the maximum-likelihood (ML) methodology or the methods of moments (MoM) were used for the parameter estimation. The method of logarithmic-cumulants (MoLC), which has been proposed for the parameter estimation for the PDF defined in R⁺, is now a very popular tool for the efficient parameter estimation for amplitude SAR images. In this article, we present a survey of some of the well-known PDFs proposed for SAR amplitude images by carrying out the parameter estimation with the MoLC method. The objective is to demonstrate that the statistical characterization of SAR images is strongly dependent upon the observed scene content. Instead of using a set of images, we carry out this study on set of object/texture categories on a larger data-base.