Research On Near-field Millimeter Wave 3D Imaging Processing Method Based On Frequency-domain Sparsity

In recent years,train stations,aerodromes and other crowded places have often suffered terrorist attacks by certain terrorists and extremist organizations,and the people have been forced to bear life and health threats also property safety losses.Traditional security check methods,such as metal detector and X-ray security detector,have some problems in terms of the types of items to be detected,the protection of passengers’ privacy,and the efficiency of security check.With the continuous development of semiconductor technology,especially millimeter-wave integrated components and processors of advanced craft,it’s gradually known that a security inspection technology based on millimeter-wave 3D imaging.And it has become a considerable research spot in the aspect of security inspection.The traditional near-field millimeter-wave 3D imaging algorithm strictly follows the Nyquist sampling law,resulting in long sampling time and large sampling data,which is not conducive to the design and implementation of hardware systems,and it also does not meet the claim of high realtime security inspection scenarios.Therefore,the study of high-resolution,high-real-time near-field millimeter wave three-dimensional imaging algorithms is the key to the largescale promotion of millimeter wave security inspection technology.Accordingly,near-field millimeter wave 3D imaging processing method based on frequency-domain sparsity was studied in this thesis.The specific content is as follows:1.For millimeter wave radar 3D imaging,the far and near field division of the radar antenna radiation and the corresponding signal model are researched,the imaging geometry and signal model of near-field millimeter wave radar are established,and the algorithm principle of near-field millimeter wave 3D imaging is derived in detail on the basis of the two-dimensional imaging principle.Then the sampling criteria and resolution of the imaging algorithm are analyzed.Finally,the correctness and effectiveness of the imaging algorithm are demonstrated through experimental simulation.2.For the problem of long sampling time and large amount of sampling data of the traditional millimeter-wave 3D imaging algorithm,MUSIC superresolution method in modern spectrum estimation is studied.On the premise of sparse sampling,the near-field millimeter wave range direction imaging problem is decomposed into the determination of target position and the estimation of scattering information.And the range imaging based on MUSIC algorithm is realized by using adjacent space smoothing technique and least square method to improve and perfect traditional MUSIC algorithm.On this basis,the 3D imaging is decomposed into a focusing problem of several 2D range sections and a near-field millimeter wave 3D imaging algorithm based on MUSIC and frequency-domain sparsity is proposed.And the validity and correctness of the proposed algorithm are verified by numerical simulation of point target imaging.3.For the problem of defocused images in the actual millimeter wave 3D imaging system work scenario due to the human body moving,3D imaging motion compensation methods are studied under the condition of sparsity,and a planar array 3D imaging motion compensation method based on PGA algorithm is proposed that do not depend on any specific error model,but make full use of the redundancy of phase error in distance plane unit.By averaging the feature points of the image,the phase error is accurately estimated to achieve the purpose of focusing the image.The availability of the proposed method is confirmed via the point target simulation experiments.