Research On Peer Group Vector Filtering Algorithm For Color Image Based On Fuzzy Logic

In the process of sampling and transmitting digital images,they are susceptible to noise from the external environment and the information in the images is destroyed.For the subsequent image recognition and feature processing work caused difficulties,affecting people to observe the image.Therefore,detecting the interference noise in the filtered image is crucial in digital image processing.The noise in the image is mainly divided into impulse noise and Gaussian noise.Usually,the impulse noise is denoised by a median filter algorithm,and the Gaussian noise is denoised by a mean filter algorithm.The requirement of image processing is to protect the edge details of the image as much as possible on the basis of denoising.However,the effects of traditional filtering algorithms do not fully satisfy this requirement.For this reason,based on the denoising problem in image processing,this paper studied two color image filtering algorithms by improving the existing filtering algorithm in the case of high noise concentration and complex composition.This paper focuses on impulse noise and improves a new color image impulse noise filtering algorithm,including noise detection and noise filtering.Before the noise detection,the noise pre-check part was added to reduce the workload of the algorithm.For possible noisy pixels,using the concept of a peer group,the samples are expanded to four directions to further detect whether there is impulsive noise.In terms of noise filtering,a method of reducing the aggregation distance was introduced to solve the problem of image detail blurring in a high-density noise environment.The determined impulse noise is removed using a modified weighted vector median filter.For mixed noise including impulse noise and Gaussian noise,this article uses fuzzylogic to filter out such noise.By referring to the concept of fuzzy set theory,the membership function is introduced to establish a fuzzy criterion,and the algorithm structure is optimized to perform noise detection on the pixels in the image.According to different noise characteristics,use the corresponding filtering algorithm.For impulse noise,the method of introducing color pairs is used to eliminate the impulse noise by using the interaction between color components.For Gaussian noise,using the weights assigned to the target pixels,the weighted mean value is calculated to replace the original pixel value,thereby filtering Gaussian noise.Noise;Pixels that are determined to be noiseless are output as they are.Simulation experiments show that the algorithm studied in this paper is simple and fast,and it achieves a good denoising function,improves the protection of image details,and enhances the edge processing effect.The project has achieved the design requirements.