AN ADAPTIVE FILTER FOR SPECKLE SUPPRESSION IN SYNTHETIC-APERTURE RADAR IMAGES

Synthetic aperture radar images are generally corrupted by speckle noi se. This arises due to the coherent nature of radar echoes used in the image formation and it is often necessary to enhance the image by spe ckle suppression before data can be used in various applications. To s uppress speckle and improve the radar image interpretability a simple filtering technique has been proposed. The filter is adaptive to the v ariance of pixel intensity in a sliding window and accordingly decides the number of nearest neighbours to the central pixel to replace its intensity with the average intensity of those nearest neighbours. The performance of the filter has been studied for speckle removal in the homogeneous areas and its edge retention capability and compared with some of the widely known speckle filters. The results show that the pr oposed filter retains edges, removes speckle noise and compares well w ith other known filters in the literature.