Non-local mean de-noising method for natural image

Abstract

The invention discloses a non-local mean de-noising method for a natural image, which belongs to the technical field of image processing and mainly solves the problem of inaccurate similarity calculation in the non-local mean de-noising of the conventional natural image. The method comprises the following processes of: (1) performing wavelet transformation on the input noise-containing natural image; (2) performing variance normalization on high-frequency information; (3) calculating the similarities of a pixel point x and a pixel point y in a search area to obtain weights of all pixel points in the search area; (4) performing weighted average on all the pixel points in the search area to obtain gray values of the corrected pixel points according to the calculated weights of all the pixel points in the search area; and (5) substituting the gray values of the corrected pixel points for the gray values of the pixel points in the input noise-containing natural image to obtain a de-noised image. The method is superior to other de-noising methods on overall performance, can keep the details of the natural image such as edges, textures and the like at the same time of better smoothing the noise, and can be used for de-noising the natural image.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a denoising method, which can be used for denoising processing of natural images. 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 an important direction in the field of image denoising. [0003] Traditional denoising methods can be roughly divided into two categories, one is the method based on the spatial domain, and the othe...

AI Technical Summary

Problems solved by technology

Due to the good performance of this method in the field of denoising, it has attracted widespread attention from many scholars since it was proposed, but it still has the following problems: 1: The complexity of the algorithm is relatively large; 2: The accuracy of weight calculation is not good; 3: The edges and details of the image are still somewhat blurred

However, wavelet transform still has the following deficiencies: 1: The sparsity of coefficients is poor in high dimensions; 2: The direction information after image decomposition is limited. In order to overcome the deficiencies of wavelet transform, Ridgelet, Curvelet, Contourlet, Brushlet, A series of new tools for image decomposition such as Bandelet

However, since the transform domain method only adjusts the shrinking threshold of the wavelet coefficients of high-frequency images, and does not process low-frequency images, the final denoising effect is not very satisfactory, and Gibbs phenomenon often occurs.

Images