Image denoising process based on Contourlet transforming

Abstract

The invention discloses an image denoising method which is based on Contourlet transformation, and belongs to the field of image processing. The realization process of the method comprises the following steps: firstly, carrying out the cycle-spinning of a noisy image so as to obtain a plurality of panning images of the noisy image; then, carrying out the Contourlet transformation of the panning images and optimizing the Contourlet transformation coefficient; then carrying out the Contourlet inverse transformation of the optimized Contourlet coefficient so as to obtain a plurality of panning images of the de-noised noisy image; carrying out the reverse cycle-spinning of the images; and then averaging the images so as to obtain the final de-noised image of the noisy image. The method which utilizes the orientation information-capturing characteristic of Contourlet has the advantages that the fine texture and the edge information of the image can be better reserved as well as the noise can be effectively suppressed by distinguishing the edges and the noise of the noisy image through the threshold method; and the distortion generated on the de-noised image can be effectively eliminated by the cycle-spinning process of the noisy image. Compared with few de-noising methods, the denoising method of the invention further has the advantages of highest PSNR value and optimal de-noising effect.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to an image denoising method based on contourlet Contourlet transformation, and relates to the application of the technology in the field of image denoising. Background technique [0002] Image noise reduction is a widely used technology in image preprocessing, its function is to improve the signal-to-noise ratio of the image and highlight the characteristics of the image. In the past two decades, people have proposed many denoising methods, which are mainly divided into the following four categories. One is the classic filter method. The processing idea is to use various smoothing functions to convolve the image, so as to achieve purpose of noise removal. These methods operate on the image through a fixed window, and do not consider the structural information around the pixel, so it is difficult to retain more image details. [0003] The second category is statistical si...

AI Technical Summary

Problems solved by technology

However, the neighborhood model introduced by this method is not realistic enough, and the calculation is complicated

[0006] Generally speaking, most of the existing methods do not take advantage of the geometric characteristics of the image. Although they have a certain effect on the removal of noise, they cannot take into account the requirements of denoising and maintaining image details at the same time.

Images