The Reserach Of Sparse Representation And Denoising Based On Cofficient Reused OMP

As the sparse representation of the signal can be expressed in fewer dictionaries of the dictionary,reducing the signal transmission and storage in the calculation of the burden of a signal processing,so it becomes a research hotspot in signal processing.Sparse decomposition stage algorithm mainly includes relaxation optimization algorithm,greedy tracking algorithm and combinatorial optimization algorithm,greedy tracking algorithm is the best real-time,but because the algorithm complexity is still high,so there is still a lot of room for improvement.In this paper,by analyzing the calculation process of greedy tracking algorithm,a more rapid coefficient reuse method is proposed,and the algorithm is applied to signal reconstruction and image denoising.The research content of the paper has the following aspects:(1)This paper analyzes the implementation process of greedy tracking algorithm,carries on the one-dimensional signal reconstruction and the image reconstruction research of OMP,StOMP,SAMP,SP,ROMP algorithm,and analyzes the advantages and disadvantages of these methods.(2)This paper proposes a sparse decomposition algorithm based on the coefficient reuse quadrature match tracing,and the algorithm is compared with other algorithms.The experimental results show that the method is more efficient in reconstructing the signal and image,and the reconstruction precision is better.(3)The new method proposed in this paper is applied to image denoising of standard image library and denoising of flame image collected by industrial field.Using the CoReOMP algorithm to solve the sparse coefficient,the use of PAU-DL algorithm to achieve a more rapid dictionary structure,the combination of the two to achieve image denoising.Based on the experimental results,the new algorithm proposed in this paper can improve the speed of signal reconstruction by 5 times compared with the traditional OMP algorithm,and the combination with dictionary learning to image denoising can achieve faster,get a satisfactory result.