eduzhai > Applied Sciences > Engineering >

A Generalized Gaussian Extension to the Rician Distribution for SAR Image Modeling

  • Save

... pages left unread,continue reading

Document pages: 20 pages

Abstract: In this paper, we present a novel statistical model, $ textit{thegeneralized-Gaussian-Rician}$ (GG-Rician) distribution, for thecharacterization of synthetic aperture radar (SAR) images. Since accuratestatistical models lead to better results in applications such as targettracking, classification, or despeckling, characterizing SAR images of variousscenes including urban, sea surface, or agricultural, is essential. Theproposed statistical model is based on the Rician distribution to model theamplitude of a complex SAR signal, the in-phase and quadrature components ofwhich are assumed to be generalized-Gaussian distributed. The proposedamplitude GG-Rician model is further extended to cover the intensity SARsignals. In the experimental analysis, the GG-Rician model is investigated foramplitude and intensity SAR images of various frequency bands and scenes incomparison to state-of-the-art statistical models that include $ mathcal{K}$,Weibull, Gamma, and Lognormal. In order to decide on the most suitable model,statistical significance analysis via Kullback-Leibler divergence andKolmogorov-Smirnov statistics are performed. The results demonstrate thesuperior performance and flexibility of the proposed model for all frequencybands and scenes and its applicability on both amplitude and intensity SARimages. The Matlab package is available atthis https URL.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...