Research On Data Processing Method For Multi-channel Through-the-wall Radar Imaging

Through-the-wall radar imaging is the technology that can detect,identify,locate and track objects within a certain range behind obstacles.In recent years,with the expansion of the application of through-the-wall radar imaging systems in the military and civilian fields,the department's requirements for through-wall radar imaging technology are also getting higher and higher.Therefore,the optimization problem of the through-the-wall radar imaging algorithm has become a popular research field.Because of the certain advantages in acquiring target related information,multi-channel through-the-wall radar imaging has attracted the public attention.Based on the establishment of a multi-channel through-the-wall radar targets signal model,this thesis proposed two compressive sensing based multi-channel through-the-wall radar imaging algorithms.The main contents of this thesis are as follows:(1)The stepped frequency through-the-wall radar system used in this thesis is introduced,and the echo signal model of the through-wall radar is established.Then the theory of compressed sensing is described in detail,and the Orthogonal Matching Pursuit(OMP)algorithm based on compressed sensing theory is introduced,which lays a foundation for subsequent research.(2)Aiming at multi-polarization through-the-wall radar measurement pattern,a Spectral Projection Gradient L1 norm(SPGL1)algorithm based on the Nonuniform Fast Fourier Transform(NUFFT)technology is proposed.The proposed algorithm avoids the enumeration calculation of the dictionary matrix by NUFFT technology,and can quickly obtain the target reconstruction result without reducing the quality of the target reconstruction.Compared with the existing greedy algorithm Look Ahead Hybrid Matching Pursuit(LAHMP),the proposed algorithm also greatly reduces the burden on computer memory.After experimental verification,the SPGL1 algorithm based on NUFFT technology improves the target clutter ratio of the target reconstruction result,and reduces the time complexity and calculating time.Compared with the LAHMP algorithm,it has higher engineering practicability.(3)Aiming at multi-view through-the-wall radar measurement pattern,a Censored Simultaneous Orthogonal Matching Pursuit Algorithm(CSOMP)based on Cross Validation(CV)is proposed.In the traditional CSOMP algorithm,the measurement noise level as the iteration termination condition belongs to the priori information,which is unknown in practice.To solve this problem,this thesis applies CV technology to CSOMP algorithm and obtains the measurement noise level adapted to the current detection scene by iteration calculation,which avoids the dependence of CSOMP algorithm on the priori information and improves the accuracy of target reconstruction results.The accuracy and practicability of the proposed algorithm are verified by simulation data and experimental data.