Non-local wiener filtering image denoising method based on singular value decomposition

Abstract

The invention discloses a non-local wiener filtering image denoising method based on singular value decomposition. The non-local wiener filtering image denoising method based on the singular value decomposition mainly solves the problem that an existing denoising method is not good in effects. The non-local wiener filtering image denoising method based on the singular value decomposition includes the achieving steps: (1) inputting a noise-contained image; (2) estimating a noise standard deviation of the noise-contained image; (3) serving an arbitrary pixel of the noise-contained image as a center, extracting an image block, and searching a similar image block of the image block in a corresponding searching window; (4) carrying out wiener filtering denoising based on the singular value decomposition on the obtained similar image block; (5) carrying out operation of the step (3) and the step (4) on all pixels of the noise-contained image; (6) reconstructing a denoising image by adoption of a block matching 3D (BM3D) gathering technology on all image blocks which are denoised; (7) judging whether iteration is finished or not, outputting the denoising image if the iteration is finished, or, transferring the denoising image serving as a noise-contained image to the step (2) to enter next iteration. The non-local wiener filtering image denoising method can effectively remove noise of a natural image which contains white Gaussian noise, and can be used for preprocessing digital images in the fields such as medical images and video multimedia.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a non-local Wiener filter image denoising method based on singular value decomposition, which can be used for digital image preprocessing in the fields of medical images, astronomical images, video multimedia and the like. Background technique [0002] With the increasing popularization of computers and digital imaging equipment, digital image processing has attracted more and more attention. However, due to the limitations of imaging equipment and imaging conditions, digital images are inevitably polluted by noise in the process of acquisition, conversion and transportation. Therefore, image denoising, as one of the basic technologies in the field of image processing, has been widely valued. Many practical noises can be approximated as Gaussian white noise, and removing Gaussian white noise in images has become a very important direction in the field of image denoising....

AI Technical Summary

Problems solved by technology

However, there are still some shortcomings in this method: because the denoising method in the first step uses a hard threshold, it is easy to cause the loss of image edges and texture details in this way; at the same time, because the denoising result of the first step is the The initial estimate of , if the initial estimate is inaccurate, it will directly affect the denoising result of the second step

The disadvantage of this method is that this method uses the PCA coefficients of all similar blocks, and many image blocks in the similar blocks are not similar enough to the image blocks to be denoised, so that the calculated coefficient variance and coefficient mean are not accurate enough, which directly affects denoising effect of Wiener filter

The disadvantage of this method is: the Wiener filter is designed using the mean and variance of each class. Since many image blocks in the class are not similar enough to the image blocks to be denoised, the calculated mean value is the same as that of LPG-PCA. and variance are not accurate enough

Images