| In reality,the acquisition of image data is likely to be polluted due to various factors.When the image is polluted by noise,the available information will be reduced,which has an important impact on the subsequent processing of the image.Therefore,it is very necessary to do a good job in image denoising.This article first introduces the research status of image denoising,and then we propose a new image denoising method in the wavelet domain.First,the image is processed by two-dimensional discrete wavelet transform(2D-DWT)to obtain a wavelet coefficient mat,rix,and then the wavelet coefficients are modeled in the wavelet domain based on the Bayesian hierarchical model framework.Specifically,organize the wavelet coefficients in the form of a quadtree structure,use the characteristics of the wavelet transform to set the prior distribution form of the wavelet coefficients,and use the position scale aggregation phenomenon existing in the wavelet coefficient layers to set the Markov tree(MT)graph structure super prior distribution.The pyramid algorithm is used to infer the model to obtain the posterior mean of the wavelet coefficients.Finally,for images with a small number of edge feat,ures,we simulate and compare the denoising effect of the algorithm in this article with other denoising methods,and find that the estimated value of MISE and LMISE obtained by the method in this paper is smaller than that of other methods,which means that the denoising effect of our method is better in both overall and local edge regions.For denoising images with complex edges,we compare the denoising effect with some of the latest denoising methods.The results show that the PSNR value obtained by our proposed method is higher than other methods,namely image quality is better after denoising,and as the image size increases,the denoising effect of our method has improved significantly.We also applied this method to real image denoising to further verify the effectiveness and superiority of this algorithm. |
PDF Full Text Request