Application Of Compressed Sensing Algorithm Based On De-noising Function In Image Processing

Since the rise of the Internet,text,pictures,videos,etc.have become an important means for people to obtain information on the Internet.Among them,pictures are the most intuitive way to obtain information.The technology of image processing is developing very rapidly,and many branches have been developed.Image restoration is one of them.Image restoration technology is approaching maturity,but the effect of restoration has room for improvement in accuracy;when a large amount of image data is encountered,if the image is compressed using a conventional algorithm,the compressed data will occupy a large amount of memory.The traditional image restoration algorithm can't remove the noise pollution of the image well,and the compressed sensing theory can make up for these shortcomings.In 2004,some famous scholars in the United States put forward the theory of compressed sensing and carried out a series of researches.No matter in domestic or foreign mathematics or engineering applications,well-known universities and multinational corporations at home and abroad have carried out corresponding research work on compressed sensing and obtained Preliminary results.The theoretical advantage is that only a small amount of data can be collected,and the original information can be completely reconstructed according to the reconstruction algorithm.Therefore,for the deficiency of traditional image restoration algorithms,the theory of compressed sensing can be well solved.Compressed sensing has been developed so far,and its application in image restoration has achieved many research results,such as base tracking(BP)reconstruction algorithm and orthogonal matching tracking(OMP)reconstruction algorithm.According to the research,the OMP algorithm has no denoising function,so that the image reconstruction effect is not very good when the image is contaminated by Gaussian random noise.The main research contents of this paper are as follows:(1)The image is sparsely represented by discrete wavelet transform.The signal is decomposed first,and the approximation signal is continuously decomposed.The signal can be decomposed into many low-resolution components,which are decomposed until the detail(high frequency)contains only a single sample.Then,the scale and displacement of the continuous wavelet transform are discretized according to the power of 2.The advantage is that with a small amount of data for storage,you can include the complete information of the original image.This application overcomes the difficulty of image compression occupying a large amount of memory when large amounts of data are encountered.(2)Improve the OMP reconstruction algorithm.Although the wavelet transform can remove some noise in the sparse process,the wavelet transform is not sensitive to Gaussian random noise,and the OMP algorithm has no denoising function.For this shortcoming,this paper improves the OMP algorithm:add one in the OMP algorithm.The algorithm step of denoising,that is,before filtering the observation vector,performs denoising processing first.Firstly,the signal is initialized,and Gaussian random noise is added to the image.Secondly,the observation vector is found and denoised.Then the observation vector is solved to obtain the signal estimation.Finally,the original information is reconstructed to obtain the reconstructed image.(3)The improved algorithm is implemented through experiments.The image is reconstructed by Wiener filtering algorithm and OMP algorithm,and then compared with the experimental results of the improved algorithm.By comparing the experimental results of different algorithms and the errors and analysis data,the conclusions are drawn.The experimental results show that the compressed sensing can improve the efficiency of the traditional image restoration algorithm and improve its shortcomings.The denoising improvement of the OMP algorithm can effectively reconstruct the image with Gaussian random noise.The greater the noise,the better the effect.The research in this paper has improved the efficiency of OMP algorithm in reconstructed images,and greatly improved the accuracy of reconstructed images.It lays a theoretical foundation for image restoration algorithms and provides some experience.Figure[29]table[1]reference[53].