Noise reduction method, device and equipment for non-local mean filtering

Abstract

The noise reduction method for non-local mean filtering comprises the following steps: acquiring one pixel in an original image as a current noise reduction pixel, and selecting a first search windowby taking the current noise reduction pixel as a center; Calculating the similarity between each pixel in the first search window and the current noise reduction pixel; Selecting the pixel with highersimilarity as the center to determine a second search window, calculating the similarity between the second search window and the current noise reduction pixel, and selecting the pixel with higher similarity as the center to determine a third search window until the boundary of the original image is searched; And selecting the pixel with the highest similarity with the current noise reduction pixel according to the calculated similarity result to perform noise reduction on the current noise reduction pixel. By calculating the similarity between the pixels in the similar window of the pixel neighborhood and the current noise reduction pixels and determining the search direction according to the pixels with the high similarity, the number of times of similarity calculation of the number oftimes of search can be greatly reduced, the complexity of noise reduction processing can be reduced, and the noise reduction efficiency can be improved.

Description

technical field [0001] The present application belongs to the field of image processing, and in particular relates to a noise reduction method, device and equipment of non-local mean filtering. Background technique [0002] The NLM (full name in English is non-local mean, and the full name in Chinese is non-local mean filter) noise reduction method can use the redundancy in the image to suppress the noise of the image while maintaining the organizational information of the image. The general process of the NLM noise reduction method is: first define two windows of different sizes around the central pixel to be processed, which are called the search window and the similarity window. The method of translation searches for similar blocks, and then obtains the filtered pixels by weighting the image block of the central pixel and the similar blocks. [0003] However, in order to effectively suppress image noise, more neighborhood image information is often introduced by increasi...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the embodiment of the present application provides a non-local mean filtering noise reduction method, device and equipment to solve the problem that the increase of the search window and the similarity window in the NLM noise reduction method in the prior art will increase the complexity of the algorithm and processing time, issues affecting the practical application of the method

Images