Kuan modified anisotropic diffusion approach for speckle filtering

Abstract

Synthetic Aperture Radar (SAR) is a coherent imagery tool used for extracting information in astronomy and meteorology. But these images are generally corrupted with a granular noise called speckle, making it difficult for extracting information. In this paper, the Anisotropic Diffusion (AD) filter is modified by incorporating the use of the Kuan filter for speckle removal. In the proposed work, the image is segmented into two regions based on the value of conduction function; the homogeneous region is processed using the Kuan filter and the other non-homogeneous region is processed with anisotropic diffusion. This modified AD filter provides better detection of weak edges and effective reconstruction of structural content with enhanced image restoration features. Further, based on the simulation results and via image quality metrics analysis; the proposed work is claimed better.