Study On Signal Reconstruction Algorithms And Design For Signal Collection Based On Comperessed Sensing

Signal sampling is a key step in the traditional signal processing. The sampling frequency must satisfy the Nyquist sampling theorem. With the bandwidth of the signal increased, the Nyquist sampling theorem gradually becomes the bottleneck of the signal processing. Recently Donoho proposed Compressed Sensing theory has proved that a signal can reconstruct the original one in the case of under-sampling.Signal reconstruction algorithm is an important component of Compressed Sensing theory. The signal reconstruction which has the robustness and low computational complexity and reconstruction effect that determines the widely used Compressed Sensing theory.This paper focuses on a greedy iterative class algorithm in-depth study. It studies of the current Compressed Sensing theory and the most common reconstruction algorithm of greedy iterative class---Matching Pursuit algorithm and Orthogonal Matching Pursuit algorithm. And this paper provides two improved methods for the OMP algorithm shortcomings:first, by using the image segmentation method to improve the rate of signal reconstruction, it obviously reduces the time of the reconstructed signal; Second, to increase the coherence between the vectors in the image signal of the reconstruction processing in order to improve the quality of reconstruction. Although this improvement is at the expense of computation time, but it still has some availability.In addition, in this paper studies the continuous signal acquisition system based on Compressed Sensing theory, the signal data which the system obtains by the OMP can accurately recover the original signal and this paper provides the simulation to achieve the whole process of the system. It proves that a continuous signal acquisition system based on Compressed Sensing theory is correct and feasible. This provides a good reference for creating hardware signal acquisition based on the Compressed Sensing theory.