Bayesian nonlocal means filter for SAR image despeckling

The nonlocal (NL) means filter as a recent denoising approach has demonstrated its empirical merit for additive Gaussian noise. In this paper, a novel Bayesian nonlocal (BNL) means filter is derived, which is adapted to the multiplicative speckle noise. The new filter is better parameterized using the Bayesian framework, in which a new statistical distance is introduced to measure the similarity of speckled patches. The proposed method is validated on real synthetic aperture radar (SAR) images through comparisons with other classical despeckling methods.