Denoising SAR images
Maria Kovaci, Dorina Isar, Alexandru Isar · 2003
The SAR images are corrupted by speckle noise. A new denoising method for this kind of images is reported in this paper. Inspired from the classical Donoho's denoising method, the procedure presented in this paper uses a new type of discrete wavelet transform, entitled Diversity Enhanced Discrete Wavelet Transform, DEDWT and a new filtering strategy. The construction of the new transform and the selection of the filtering strategy applied in the wavelet transform domain are inspired from methods used in communications. For the construction of the new wavelet transform used the principle of diversity enhancement and for the filtering in the wavelet transform domain used the principle of packet's filtering. Some simulation results prove the performances of the proposed denoising method.