Millimeter Wave Human Body Three-Dimensional Imaging And Target Detection

The international and domestic public security situation has become increasingly severe in recent years,which has put forward higher requirements for human security inspections in densely populated public safety places,such as airports,railway stations and subways.Conventional security inspection methods such as metal detectors and X-ray security inspection machines cannot be used safely and effectively for human body inspection due to their respective limitations.Therefore,it is necessary and urgent to develop modern body inspection technology.The millimeter wave security inspection technology has the inherent advantages of strong penetrating ability,small radiation to the human body and high resolution,and has become an important alternative means for modern human security inspection.On the basis of millimeter wave security imaging,combined with dangerous object detection technology to achieve detection of hidden objects in human body,it is a safe and effective body inspection program,and has broad application prospects.The real-time performance and accuracy of human body imaging and detection is the focus of the millimeter-wave body inspection system in practical applications.For different imaging mechanisms,the corresponding real-time imaging and detection algorithms are the hotspot of current research.This paper combines the widely used cylindrical scanning millimeter wave three-dimensional imaging human body security inspection system to study three-dimensional scene reconstruction algorithm,millimeter wave human body three-dimensional real-time imaging,millimeter wave image target detection algorithm,the main contents include:(1)About characteristics of cylindrical scanning millimeter wave human body three-dimensional imaging system,the more mature three-dimensional wave number domain imaging algorithm and the mathematical model of high-precision SAR focusing algorithm are introduced.Based on the wave number domain algorithm and the backprojection algorithm,combined with the low complexity of the ?K algorithm and the precise focusing of the BP algorithm,a human body 3D imaging algorithm based on the ?K algorithm and the BP algorithm(?K-BP algorithm)is proposed.(2)Based on the theoretical derivation of the millimeter-wave human body three-dimensional imaging algorithm,the correctness of the theoretical derivation of the three-dimensional human imaging algorithm is verified by simulation data,and the performance of different imaging algorithms is analyzed and verified.Then,through the comparison of the imaging results,the advantages and disadvantages of several commonly used millimeter-wave three-dimensional imaging algorithms are analyzed,and the application of the three-dimensional imaging algorithm in engineering is verified by imaging the echo data collected by the actual cylindrical scanning millimeter-wave three-dimensional imaging system.(3)Aiming at the easy parallel design of BP algorithm in ?K-BP algorithm,the real-time implementation of human body 3D imaging based on CUDA parallel programming technology is studied.The realization scheme of the three-dimensional imaging algorithm combining ?K and BP algorithm on the GPU general parallel computing platform is given.The engineering realization design of the algorithm is realized and the purpose of human body three-dimensional real-time imaging is achieved.(4)Based on the three-dimensional imaging of millimeter-wave human body,the detection of hidden objects in human body is studied.The deep learning neural network optical image recognition algorithm,such as the RCNN class algorithm,is applied to human body millimeter wave image recognition detection.On this basis,aiming at the characteristics of active cylindrical scanning millimeter wave human security image,the idea of selecting the area to be inspected is improved,and a new detection method for the hidden target of human body are given.