Matching Pursuit Algorithm In Compressed Sensing And Its Application

With the rapid development of information technology,there are massive amounts of information in various application fields.The traditional signal processing method is encountering its bottleneck period.Compressed Sensing is a new developed signal processing method and it is different from the traditional method.It can recover original data exactly from a small number of sampling data via reconstruction algorithm.Therefore,this paper mainly studies the matching pursuit reconstruction algorithm and its improved algorithm in Compressed Sensing.It is applied in the reconstruction of one-dimensional logging data and the reconstruction of two-dimensional remote sensing images.The main work or innovations are as follows:(1)The research and simulation on matching pursuit algorithm.Study the performance of a variety of matching pursuit algorithm and an improved sparsity adaptive matching pursuit algorithm based on Compressed Sensing.The matching pursuit algorithm is used in one-dimensional Gaussian random sparse signal reconstruction and binary random sparse signal reconstruction.The simulation results of performance test show that the MP type algorithm can reconstruct the original signal with extremely high probability at the sampling rate of 0.5.The simulation results of performance comparison show that,in the same condition,SP,SAMP and improved SAMP algorithm have a similar reconstruction performance,and is superior to other algorithms.(2)The reconstruction of logging data based on Compressed Sensing.In order to alleviate the problem in processing massive logging data,the reconstruction of logging data based on Compressed Sensing is studied.A variety of matching pursuit type algorithm is used in logging data reconstruction of a typical well in Xinjiang.And the experiment is based on K-SVD over-complete dictionary.The results show that the matching pursuit type algorithm is effective and feasible in logging data reconstruction.By contrasting the reconstruction effects,it shows that the improved SAMP algorithm has a better performance in logging data reconstruction.(3)The reconstruction of GaoFen-1 remote sensing images based on Compressed Sensing.In order to solve the shortages of traditional method in dealing with the massive remote sensing image data,the reconstruction of GaoFen-1 remote sensing images based on Compressed Sensing is studied.Matching pursuit type algorithm is used in GaoFen-1 remote sensing images reconstruction based on K-SVD over-complete dictionary.The results show that matching pursuit type algorithm is effective and feasible in remote sensing images reconstruction.Under the same reconstruction algorithm,the reconstruction effect based on K-SVD over-complete dictionary is better than that based on DCT over-complete dictionary.Under the same sparse dictionary,the improved SAMP algorithm has a better performance with less running time.