Adaptive parametric estimator for complex valued images

Abstract

Synthetic Aperture Radar (SAR) images are affected by speckle which has a multiplicative nature. It is a granular noise which appears due to coherent interference of radar waves reflected from elementary scatterers. The presence of speckle greatly affects the analytical ease in terms of image classification and segmentation. Despeckling filters such as Lee, Frost, Boxcar and other have been proposed to deal with the phenomenon of speckle. However, the major concern for such filtering approaches is the preservation of features like edges and point targets while filtering the homogeneous areas. In this paper, a new approach is proposed which deals with this issue by using a statistical method that filters the different image features separately. Thus, it provides a greater level of control on the filtering process. The performance of the proposed filter is analyzed using AIRSAR C-band dataset. The results show that the proposed filter gives a good performance in terms of despeckling the homogeneous region whilst preserving the edges and point targets.