Weighted Bayesian based speckle de-noising of SAR image in contourlet domain
Mohamad Kiani Anbouhi, Sedigheh Ghofrani · 2014
The main problem of applying Bayesian Shrinkage in transform domain such as Contourlet transform (CT) or wavelet transform (WT) is finding the optimum threshold values. In this paper, we show that the Contourlet coefficients are affected by noise differently. It means, some Contourlet coefficients belong to the specific sub-bands are more robust against noise. We use this new found property and define the noise efficiency factors in order to determine the optimum threshold values and develop our proposed method based on weighted Bayesian Shrinkage in Contourlet domain. Obtaining the optimum threshold values remove more speckle noise of SAR image and preserves the quality of image as well. Four objective assessment parameters are computed in order to evaluate our proposed method in comparison with two other ordinary de-noising approaches.