**6. Conclusion**

This book chapter has presented the proposed methods for despeckling a synthetic and real SAR images using second-generation wavelets. The Bayesian approach in incorporated into second generation wavelets using the wavelet domain. The prior and likelihood pdfs are modeled using GGMRF and Gaussian distribution. The second order Bayesian inference is used to better estimate model parameters and to find the best values possible. The evidence has been simplified and approximated using the Hessian approach. The experimental results have shown that the despeckling of real SAR images using second-generation wavelets is comparable with the dyadic wavelet-based despeckling algorithm (Gleich & Datcu, 2006). Moreover, information extracted using the contourlet domain gives good results using synthetic as well as real SAR data. Unfortunately, the contourlet-based despeckling introduces lines, which are consequences of cutting low-frequency components in the subband decomposition, which can be corrected by introducing a new filter or by post-processing step.
