Performance Measure of Despeckling of SAR Image using Multiwavelets
Jawed Qumar, Motorola India Bangalore, Sumana Gupta · 2009
Abstract — Synthetic aperture radars (SAR) are coherent imag-ing systems that produce complex-valued images of the ground. SAR images are corrupted by multiplicative noise (speckle Noise). It is due to coherent processing where multiscatterers present in a resolution cell. The presence of speckle in an image reduces the resolution of the image and the detectability of the target. The Multiwavelet transformations are useful for speckle reduction through its sub-band images. The speckle noise variances are estimated first using which donoho thereshold is computed. Then transform coefficients are pruned to reduce the speckle noise. The proposed method shows great promise to reduce the speckle noise and evaluates the performance measure for dif-ferent Multiwavelets. In this paper several thresholding scheme are used with different Multiwavelets and Prefilters. I.