Recently a family of iterative greedy algorithms have received extensive application in compressed sensing (CS) due to their fast reconstruction and low reconstruction complexity. In this paper, the basic theory of CS is first introduced and then we put emphasis on the main greedy algorithms for reconstruction, which include MP, OMP, IBOOMP, StOMP, SP, ROMP, CoSaMP and so on and provide their mathematical frameworks, respectively. Next, we classify all the algorithms according to the strategy of element selection and the update of the residual error. Under the condition of restricted isometry constant, further discussion on the performance of reconstruction algorithms such as running time, reconstruction stability and so on is presented. Last, the reconstruction results from simulated experiments further show the performance of all algorithms. And from those results we also acquire the relationship among the performance of the algorithms, the sparsity of signals to be reconstructed and the number of measurements, which lays a good basis for proposing new and better algorithms.
国家科技期刊平台
NSTL
万方数据
维普
