Research On Image Denoising Method Of Microorganisms In Sewage

At present,the activated sludge process is the most effective method in the municipal sewage treatment system.Microorganism plays an important role in indicating the quality of sewage treatment.Therefore,the detection technology of sewage microorganism based on digital image processing has been widely used,and the denoising of microorganism image is one of the key links.In this paper,based on the microorganism image collected in the urban sewage treatment system,through a large number of research literature and data,the denoising algorithm for different algorithms of sewage microbial image are studied and improved.The main contents are as follows:This paper introduces the basic theoretical knowledge of image denoising,including the types and characteristics of noise in image,quality evaluation of denoising image,approximation of low-rank matrix of image and sparse coding theory of image.In order to solve the problem of color distortion and fuzzy easily generated by color microbial image denoising,a new denoising algorithm based on bilateral weighted pseudo norm is designed.Based on the multi-channel characteristics of color image,two weight matrices(2₁ and(2₂ are introduced.They can solve the noise difference between different channels according to different noise standard deviation,and improve the color distortion and fuzzy state.The algorithm approximates the rank function with pseudo norm,which can automatically correct the proportion of different singular values in the rank function minimization model according to the singular value of the rank function,and effectively improve the signal-to-noise ratio of the denoised image.The experimental results show that the algorithm can keep the internal structure of microorganism better,the average peak signal-to-noise ratio of image is about 6d B higher than other contrast algorithms,and the structural similarity is about 19%higher on average.In view of the Poisson noise in microbial image,it is easy to produce step effect and edge blur after denoising.A sparse Poisson denoising algorithm based on image block clustering is proposed,which can denoise Poisson noise directly.According to the intrinsic correlation of microbial images,the width of the blocks is determined automatically by the similarity of adjacent images after the block is segmented;in addition,the best clustering number K value is determined by elbow method,and then the Poisson K-means method is used to cluster.Principal component analysis is used to represent the clustered image block matrix in dictionary.Sparsity constraint is added to update the dictionary coefficients.Only a few important dictionary elements are used to represent each image block.The experimental results show that the algorithm improves the edge blur of microbial targets in the image,reduces the step effect,and retains the image details as much as possible.The average peak signal-to-noise ratio is about 7d B higher than other contrast algorithms,and the structural similarity is about22%higher on average.Aiming at the problems of mixed noise in color microorganism image,which is easy to produce artifact and excessive smoothing in denoising,this paper proposes a trilateral weighted sparse coding denoising algorithm to denoise the mixed noise of additive Gaussian white noise and impulse noise.Specifically,two weight matrices are introduced into the data fidelity term to adaptively describe the noise statistics in different channels and image patches,and another weight matrix is introduced into the regularization term to make better use of the data sparsity prior.Experimental results show that the proposed algorithm can reduce the generation of artifacts,improve the over-smoothing condition,preserve the edge and internal structure of microorganism image well,and the peak signal-to-noise ratio is about 7d B higher than other algorithms on average,and the structural similarity is about 17%higher on average.