Sea clutter modeling in presence of thermal noise using beta-prime texture distribution

Modeling of high resolution sea radar clutter using compound-Gaussian model with additive thermal noise is proposed in this communication. To do this, the Beta-Prime distribution is selected as a texture component. Then, the overall PDF (Probability Distributed Function) is given in integral form in terms of three parameters. As the closed form methods to estimate these parameters cannot be achieved, we resort in this work to use the parametric curve fitting estimate approach based on the N-M (Nelder-Mead) algorithm to get the optimal estimates of the parameters. It is shown, with reference to real IPIX high resolution radar clutter that the compound Beta-Prime distribution can yield tighter fits compared to the lf-plus-noise and the Pareto-plus-noise models in most case.