Research On Application Method Of Compressed Sensing Theory In Space Optical Remote Sensing Imaging System

With the increasing imaging scale and resolution requirements of space remote sensing imaging systems,it poses a huge challenge to traditional imaging systems.The increase in imaging resolution requires a larger pixel array size and a smaller pixel size,and the massive amount of data generated also puts tremendous pressure on the storage and transmission of data.Compressed sensing theory breaks through the limitation of Shannon-Nyquist sampling theorem in traditional imaging systems.Based on the sparse characteristics of signals,it can reconstruct the original signal accurately from a small number of projections with high probability,which provides a new theoretical basis for the design of high-resolution imaging systems.In this paper,from the perspective of compressed sensing theory model and system implementation,the compressed sensing imaging system and method for space optical remote sensing application are studied.The improvement of theoretical model,hardware implementation and imaging strategy are discussed in depth.Firstly,a parallel complementary compressed sensing imaging system is proposed,and the mathematical model of image block parallel processing is established.Compared with the traditional block compressed sensing system,the sampling efficiency is improved and the imaging quality is better.According to the application requirements in the field of space optical remote sensing imaging,the push broom image motion model of the system is established.The effect of the image motion to the quality of restored image is quantitatively analyzed.Secondly,based on the compressed sensing theory and the working mode of TDI CCD,a compression coding imaging method based on TDI CCD charge transfer and shutter random flash exposure is proposed to improve the image resolution in the one-dimensional direction of charge transfer.Based on this method,a dual detector orthogonal super-resolution imaging system is proposed,which improves the resolution of image in two-dimensional direction.A two-dimensional image fusion method based on greedy algorithm and proportional fusion algorithm is proposed.Theoretical modeling and simulation results show the effectiveness of the imaging method and system,which provides a new method for system implementation of super-resolution imaging based on compressed sensing theory.Thirdly,an improved model of the compressed sensing theory model is proposed,and the hardware implementation scheme and system workflow are designed according to the model requirements.The simulation results show that compared with the block compressed sensing system,the proposed system can significantly improve the image quality under the same parameters and eliminate the block effect of the block compressed sensing system.On the basis of the proposed system,a weighted iterative restoration strategy based on frequency prediction is proposed.By adaptively adjusting the reconstruction process,the resource allocation of the reconstruction process is optimized,and the image quality is further improved.Finally,according to the saliency characteristics of different regions in remote sensing images,an adaptive block compression sensing system is proposed,which can obtain the prior information of the target scene by pre-sampling first.On the one hand,it is used to guide the adaptive sampling rate allocation between blocks in the sensing stage,on the other hand,it is used to generate the weight sparse coefficient matrix in the recovery stage,so as to improve the efficiency of image restoration.Optimized combination application strategy makes the image quality of the proposed system significantly improved compared with the traditional system.At the same time,a multi-angle image restoration strategy is also proposed,which greatly improves the quality of image restoration.In this paper,the space optical remote sensing imaging system and method based on compressed sensing theory are studied in depth,and the theoretical model and system implementation of compressed sensing are explored effectively,which promotes the application of compressed sensing theory in the field of space optical remote sensing,and also has a certain guiding significance.