Edge Preservingreserving Despeckling Algorithm for Remote Sensing Synthetic Aperture Radar Images using Multi -Scale Curvelet Transform and Adaptive Thresholding

Curvelet transform is a new kind of multi-scale analysis algorithm which is veryuseful for SAR image processing. This transform is considered as geometric transform which is developed to overcame the vital limitations of wavelet and gabor transforms in edges and curves detection. The curvelet transform can be used with very high precision a better segmentation and target detection in SAR satellite images. In this paper, the main goal is to focus on the despeckling of SAR images which is the most important stage in the processing of SAR images. Because of excellent results of curvelet transform in the analysis of curved shape edges and its high accuracy, this transform is useful to approximate and describe the dispersals and orientations in comparison with the wavelet. In this paper, using fast discrete curvelet transform (FDCT) based on wrest and using an adaptive thresholding, a new despeckling algorithm is developed for SAR images; So that a comparison will be done on both fast curvelet based on USFFT and based on Wrapping. Using the results, a new method to take advantage of this kind of images is introduced. Then, the Fast Discrete Curvelet transform (FDCT) is used for noise reduction of SAR images and the Curvelet coefficients are decreased. Then, using the the proposed algorithm the curvelet coefficients can be extracted. The curvelet coefficients are affected by soft thresholding and this leads to reduce or eliminate the noise in SAR images. Finally, the proposed algorithm is simulated, and the results will be discussed. Simulation results demonstrate that the proposed algorithm has better performance than the other similar methods.