SAR image despeckling based on Gaussian scale mixtures model in stationary wavelet domain
Chi Zhang · Journal of Heilongjiang Institute of Science and Technology · 2010
Aimed at overcoming blurring edge and distortion revealed by traditional methods used in image processing,this paper proposes the combination of stationary wavelet transform with Gaussian scale mixtures model to develop the neighborhood model based on stationary wavelet transform coefficients and the reduction of the speckle noise by using Bayes least square estimation.The paper ends with the objective assessment of filtering results of the various methods by using evaluation parameters.Experimental results show that this method exhibits the speckle noise-reducing performance better than the spatial filters and standard wavelet transform.