Research On Image Restoration Algorithm Based On Compressed Sensing

Compressed Sensing(CS)is a new signal sampling theory in recent years.It breaks through the traditional Nyquist sampling theorem and subverts the traditional sampling method.For the sparse or compressible signal,it's compressed and sampled,and then the reconstruction algorithm is used to reconstruct the signal accurately.The technology of digital image processing is an important research filed ant it has been widely used in our life.With the development of digital technology,research on image reconstruction and image denoising will be more and more.In particular,the use of new technologies in those has become the inevitable trend.In this thesis,the image denoising and image reconstruction are studied.The main work of this thesis is as follows:An introduction to the compressed sensing.The experiment is carried out to verify the effectiveness and feasibility of it.An introduction to the image denoising and the qualities of image.And finally,the simulation experiments of image denoising are carried out.This thesis introduces the Orthogonal Matching Pursuit(OMP)algorithm,and found that the residuals and each column of the sensing matrix have an inner product,which can result in a large amount of computation and long time consuming.To solve this problem,an improved OMP algorithm based on dynamic threshold is proposed on the basis of OMP algorithm.After the inner product operation of residual and the column of the sensing matrix,we can get an average and take it as a threshold value.The column of the sensing matrix below the threshold is not involved in the next operation.The experimental results show that,compare with the OMP algorithm,the improved OMP algorithm and the OMP algorithm have the same reconstruction effect,at the same time,the improved OMP algorithm can greatly reduce the computational complexity and the running time.An introduction to wavelet transform and wavelet threshold image denoising.The hard threshold denoising and soft threshold image denoising are analyzed.Based on the analysis,an improved threshold image denoising method is proposed.Below the threshold,the wavelet coefficients are multiplied by a number on the 0-1,so to retain the effective image information.Higher than the threshold,the wavelet coefficient minus a number.The experimental results show that the improved method is feasibleand effective,and the denoising effect is better than the hard threshold denoising and soft threshold denoising method.Finally,we combine the improved OMP algorithm with the improved threshold denoising method,and the experimental results show that the improved threshold method has better denoising effect.