2013/04/16 by Leonardo A. B. Tôrres, Sidnei J. S. Sant’Anna, Torres, Leonardo +4 · 1 citation
Engineering · Mathematics · #Applications (stat.AP) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Graphics (cs.GR) #Information Theory (cs.IT) #Machine Learning (stat.ML) #Remote-Sensing Image Classification #Statistical and numerical algorithms #Synthetic Aperture Radar (SAR) Applications and Techniques
paper · pdf · doi:10.48550/arxiv.1304.4634
openalex publication_date 2013/04/16 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
This paper presents a technique for reducing speckle in Polarimetric\nSynthetic Aperture Radar (PolSAR) imagery using Nonlocal Means and a\nstatistical test based on stochastic divergences. The main objective is to\nselect homogeneous pixels in the filtering area through statistical tests\nbetween distributions. This proposal uses the complex Wishart model to describe\nPolSAR data, but the technique can be extended to other models. The weights of\nthe location-variant linear filter are function of the p-values of tests which\nverify the hypothesis that two samples come from the same distribution and,\ntherefore, can be used to compute a local mean. The test stems from the family\nof (h-phi) divergences which originated in Information Theory. This novel\ntechnique was compared with the Boxcar, Refined Lee and IDAN filters. Image\nquality assessment methods on simulated and real data are employed to validate\nthe performance of this approach. We show that the proposed filter also\nenhances the polarimetric entropy and preserves the scattering information of\nthe targets.\n