A structure-aware despeckling method of SAR image

Recently, Synthetic Aperture Radar (SAR) has been widely applied in ocean remote sensing, since its charming advantages. However, the subsequent understanding of SAR images is difficult because of the existence of speckles. Traditional denoising methods fail to remove speckles while preserving edges. Local anisotropy degree derived from nonlinear structure tensor was introduced to distinguish structures from speckles in this paper. A structure-aware despeckling algorithm was proposed under the tensor based weighted least squares framework. Experiments demonstrated the effectiveness and robustness of the proposed method.