SAR Image De-noising Using ICA Basis Enhancement and Separation

Proposed is a novel theory for SAR (synthetic aperture radar) image de-noising based on ICA (independent component analysis) basis images. Firstly, the basis images and the code matrix of the original image are obtained using ICA algorithm. Then, pointwise Holder exponent of each basis is computed as a cost criterion for basis enhancement, and then the enhanced basis images are classified into two sets according to a separation rule which means the clean basis can be separated from the original basis. These two procedures are the key points for speckle reduction. Finally, the clean image is obtained by reconstruction on the clean basis and original code matrix. As the results show, the reconstructed image achieves a better visual perception compared to those obtained by other traditional techniques, and a lower value of SD/M (ratio of standard deviation to mean) factor has been calculated demonstrating better image quality.