Wavelet Based Speckle Filtering of the SAR Images

Abstract

Synthetic Aperture Radar (SAR) images are corrupted by speckle noise due to random interference of electromagnetic waves. The speckle degrades the quality of the image and makes difficult to interpret, analysis and classification of the SAR images. This paper proposes a method that reduces the speckle and preserves the features. The features are preserved by using scale-space correlation between the scales. The results show that the proposed method is better than widely used filters based on the spatial domain, such as Lee, Kuan, Frost, Ehfrost, Median, Gamma filters in terms of feature preservation. Moreover the proposed method achieves a wide range of balances between speckle reduction and feature preservation, and thus is applicable in different applications such as road detection, bridge, and ribbon like structure. Furthermore, the proposed method does not require prior modeling of either the image or noise statistics. It uses the variance of the detail wavelets coefficients to estimate noise variance.