Research On Image Reconstruction Of Millimeter Wave Imaging Based On Compressed Sensing

Millimeter wave imaging can be used for security,aircraft blind drop,battlefield environment imaging and other fields,is the current research focus.At present,the passive millimeter wave imaging system uses multi-channel scanning structure to achieve rapid imaging,but the channels are more difficult to debug and system cost is higher.In this paper,a single channel passive millimeter-wave imaging system based on CS is carried out,which mainly focuses on coding and image reconstruction.The main works are as follows.(1)The design of millimeter-wave single channel coding imaging scheme is completed.On the basis of analyzing the theory of passive millimeter wave imaging,this paper analyzes,computes and designs the optical path of single-channel code imaging.Based on the CS theory,a mathematical analysis model of millimeter-wave single-channel coding imaging is established.(2)The design of millimeter-wave single channel imaging coding template based on parity check matrix of LDPC codes is completed.The analysis of several typical measurement matrices,such as Hadamard matrix and cyclic S matrix,this paper proposes a measurement matrix based on LDPC code parity check matrix based on PEG-quasi-cyclic construction method,which is verified by simulation results.And it has advantages of excellent reconstruction performance,strong structure and recyclability compared with other methods.At the same time,the hardware design based on the LDPC code parity check matrix as the coding template is completed.(3)The reconstruction of image by single-channel coding is realized by TV-structure clustering super-resolution image reconstruction algorithm.Based on the deep research of a variety of CS image reconstruction methods,this paper focuses on super-resolution reconstruction of sparse representation of millimeter-wave images,compared with sparse representation of K-SVD dictionary and reconstruction of structural clustering representation,the performance of the proposed method is reconstructed by the superposition method of structure clustering representation,and the algorithm is optimized by TV.The simulation results verify the effectiveness of the method.