SWOMP RECONSTRUCTION ALGORITHM BASED ON MEAN THRESHOLD AND BACKTRACKING STRATEGY

In order to improve the reconstruction accuracy and running speed of the staged weak selection orthogonal matching pursuit(SWOMP)algorithm in the compressed sensing reconstruction algorithm,a SWOMP algorithm based on the mean threshold and backtracking strategy is proposed.The algorithm used the mean strategy to adaptively select atoms,which improved the accuracy of atomic screening.We used the backtracking strategy to perform secondary screening on the selected atoms,and optimized the support set to improve the reconstruction accuracy of the algorithm.The matrix design was simplified to reduce the iteration times of the algorithm,which increased the running speed of the algorithm.Simulation experiments show that the reconstruction performance of this algorithm for one-dimensional random signals and two-dimensional image signals is significantly better than other similar algorithms,and it has the characteristics of high reconstruction accuracy and less time consumption.