SAR image despeckling based on G0 distribution and nonlocal TV regularization
Xiangli Nie, Hong Qiao, Bo Zhang, Suiwu Zheng · 2013
In this paper, we propose a new model for synthetic aperture radar (SAR) image despeckling based on the G0statistical distribution and nonlocal total variation regularization. By taking the distribution of the backscatter into account, a new data fidelity term is derived by the maximum a posteriori Bayesian rule. Combining the new fidelity term with the nonlocal total variation regularization gives a new variational model for SAR image despeckling. The primal-dual algorithm framework is then used to solve the new variational problem. Experimental results on real SAR images demonstrate the validity of the proposed method.