Wavelet transform based technique for speckle noise suppression and data compression for SAR images

A compression system based on wavelet transform zero-tree coding has been applied after suppressing the speckle noise in synthetic aperture radar (SAR) imagery. We have performed normalization and shrinking of wavelet coefficients of SAR images to remove the speckles from the SAR imagery and then apply wavelet based set partitioning in hierarchical trees (SPHIT) algorithm for image compression which further improves the quality. Since radar images contain multiplicative speckle noise, the normalization technique is used to convert multiplicative noise into additive noise, and then remove it by shrinkage of wavelet coefficients. The normalization is done with respect to coarse scale (low frequency) wavelet coefficients and applied to all finer scale coefficients (high frequency) spatially related with coarse scale coefficients. This works well since wavelet coefficients are modulated by the multiplicative character of the speckle in a manner that is proportional to the target mean back scattering coefficient. Four types of test images have been selected for the demonstration of results and excellent quality reconstruction are obtained at data rates as low as 0.5 bpp for detected imageries.