Bayesian speckle noise reduction using the discrete wavelet transform
Zhaohui Zeng, Iain Cumming · 1998
Speckle noise is an inherent property of SAR images. The discrete wavelet transform results in sub-band decompositions of SAR images which have Laplacian point statistics. The authors propose a Bayesian estimator to perform speckle reduction. They evaluate the speckle reduction results using two measurements: standard-deviation-to-mean ratio and target-to-clutter ratio. They show that good speckle reduction results can be achieved while maintaining the significant information of the SAR image.