Three-dimensional matching coordination filtering image de-noising method

Abstract

The invention discloses a three-dimensional matching coordination filtering image de-noising method based on a transforming domain. The method comprises the following steps of 1, performing gray scaleconversion on an acquired image and performing noising processing; 2, obtaining pre-estimated data of the image by means of three-dimensional matching filtering; 3, performing wavelet decomposition conversion for extracting a high-frequency part in the pre-estimated image obtained in the step 2 and performing filtering, for avoiding edge fuzzy, introducing a Laplace Gaussian algorithm for constructing a new operator and introducing into a diffusion model, and finally performing wavelet reconstruction for obtaining final approximation of the original image; and 4, detecting the edge through the Canny operator and calculating the peak value signal-to-noise ratio and structure similarity of each model for analyzing the effectiveness and feasibility of the method. The three-dimensional matching coordination filtering image de-noising method has advantages of excellent performance, whole inner information protection, reasonable operation speed and benefit for actual application.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a three-dimensional matching harmonic filter image denoising method based on transform domain. Background technique [0002] Image has become the most commonly used information carrier in human activities at this stage, and the research on image processing processes such as edge detection, feature extraction, and pattern recognition has become more and more popular among scholars. However, it is very easy to introduce noise in the process of image acquisition, compression or transmission, which will affect the subsequent work. Therefore, noise reduction processing is particularly important in the image processing process. Most of the denoising algorithms at this stage are derived from the fields of probability statistics theory, fuzzy theory, non-parametric estimation theory and so on. Due to the lack of prior information, the denoising problem is often ill-conditioned...

AI Technical Summary

Problems solved by technology

[0003] At present, there are many studies on denoising methods. Many current image denoising algorithms lack the analysis of the overall structure and the amount of calculation is too large. Therefore, it is urgent to study the relevant content and get a perfect conclusion. Provides an effective method for a better balance between the protection of structural information and computational complexity

Images