Model-Based Multiresolution Restoration of Speckle Images: Application to Radar Imaging
Pierre Moulin · 2005
We consider the problem of restoring images corrupted by speckle noise with iid-exponential statistics. This model occurs in a variety of coherent imaging problems including diffuse-target radar imaging. Our estimation approach is obtained by statistical inference on the wavelet coefficients of the logarithm of the image. Under a large-sample approximation, the inference problem takes a very simple form. Estimates can be obtained under various noise/resolution tradeoffs.