Underwater Infrared Image Denoising Based on Un-decimated Wavelet Transform Bayesian Estimation
Jun Zhou · Ocean Technology · 2010
Considering the characteristic of shift invariance and redundancy of un-decimated wavelet transform,the model for wavelet coefficients based on Bayesian estimation and Generalized Gaussian distribution(GGD) is built,finally,constructs MapShrink subband adaptive image denoising algorithm and thresholding function based on Laplacian distribution.The algorithm makes 2J extraction for HH1 subband coefficients,not only estimates noise variance but also effectively restrains Pseudo-Gibbs phenomenon.Simulation results indicate that the proposed underwater infrared image denoising method gives better SNR and MSE performance than traditional Bayesian denoising methods.