Dice Coefficient Based Image Reconstruction Algorithms Of The Compressed Sensing

As the rapid development of information technology in recent years, traditional signal processing in accordance with the Shannon-Nyquist sampling theorem, it requires only when the sampling frequency is not less than twice the maximum bandwidth of the signal to recover the original signal without distortion. Nowadays digital image is complex and huge, continuing to use the Nyquist sampling theorem for data processing, will bring information transmission, storage and processing tremendous pressure and it is difficult to meet the requirements for data processing. The emergence of compressed sensing theory breaks the shackles of the Nyquist sampling theorem, making new way of signal processing, process signal sampling and compression at the same time. CS theory bring signal processing a major change.By studying the developing of compressed sensing theory, as well as details of the basic compressed sensing theory flow path, research common signal reconstruction algorithm and the current defect, through studying these issues this paper has the main job as follows:① Analysis signal sparse representation of the compressed sensing theory, design of observation matrix and signal reconstruction algorithm. The main study included some major greedy algorithm like Matching Pursuit, Orthogonal Matching Pursuit, Stagewise Orthogonal Matching Pursuit, Compressed Sampling Matching Pursuit, study and analysis the advantages and disadvantages of algorithm implementation steps and signal reconstruction result.②Analysis best atomic matching processes of Orthogonal Matching Pursuit algorithm, find defects in inner produce rule method, combined with Dice coefficient as a new atomic matching rule, come up with a new algorithm based on Dice coefficient named Dice-OMP, to solve inconvenient to select the appropriate atoms support constituting problem in OMP algorithm.③Analysis Dice-OMP signal reconstruction algorithm, from the reconstruction success rate, signal reconstruction error and signal reconstruction time of algorithm to do matlab simulation comparison to prove the practicality of Dice-OMP algorithm.④Simulated data experiments for two-dimensional image reconstruction to demonstrate the feasibility of Dice-OMP algorithm, it has better reconstruction results, lower reconstruction error rate at different sampling rates, which can be used as an improved edition of OMP.