Image Reconstruction Algorithm Based On Compressed Sensing And Its Application

With the rapid development in the field of signal processing in recent years,novel sampling theorems have been put forward one after another.The traditional sampling theorem is no longer suitable for the contemporary society.The demand for data acquisition in the new era has gradually become huge.The constantly updated data In this context,the compressed sensing theory has evolved.It takes a certain amount of observations and reconstructs the initial signals from a small number of samples.Generally speaking,this is a process of solving undetermined equations.Based on the observed values of Y,a series of algorithms such as convex optimization algorithm or greedy algorithm are used to find the solutions of the underdetermined equations,so that the initial reconstruction with high precision Signal X.In summary,this paper will innovate on the reconstruction algorithm and propose an lp image reconstruction algorithm.In addition,based on the Co Sa MP reconstruction algorithm,it is optimized and improved,and an FR-Co Sa MP reconstruction algorithm is proposed.The main work includes:?1?The performance and parameters of convex optimization algorithms such as l₀,l₁ norm are analyzed.The two-dimensional image is reconstructed by using Matlab software.Based on the different characteristics of convex optimization algorithm,and according to the problems of low precision of image reconstruction and low efficiency of the algorithm,Based on the lp norm image reconstruction algorithm,the mathematical model and algorithm flow of the algorithm are analyzed in detail.Finally the Matlab simulation test is used to obtain a qualitative leap in the image reconstruction accuracy and reconstruction efficiency.?2?The Co Sa MP algorithm,one of the greedy algorithms,is studied and analyzed.Based on this,an improved FR-Co Sa MP reconstruction algorithm is proposed.The improved algorithm is based on the Co Sa MP algorithm,and the iterative calculation method is optimized to improve the reconstruction quality of the image signal.Then,after comparing the OMP,St OMP and FR-Co Sa MP simulation images,the optimized FR-Co Sa MP has a significant improvement in algorithm efficiency.?3?Due to the development of technology,the reconstruction on two-dimensional images has been unable to meet the current development needs.It was decided to applythe image reconstruction optimization algorithm to three-dimensional video detection and propose a new spatial domain video compressive sensing model.The director of the formula is trained and the target is reconstructed and tracked with the moving object as the goal.According to the results of the operation on the three-dimensional video,under certain conditions,good reconstruction and tracking of the three-dimensional video can be achieved.