Research On Active Millimeter-wave Imaging Algorithm For Security Insepection

In recent years,because of the terrorist attacks occurring in public places such as airport as well as Millimeter-Wave(MMW)possessing penetrability and non-ionization,a new safe and effective Millimeter-Wave(MMW)security imaging technology emerges as the times require.Based on the MMW security check imaging,this thesis studies the active near-field MMW imaging algorithm,and on this basis,this thesis reconstructs imaging of sparse samples data by using the compression perception theory.Firstly,in this thesis,an active near-field MMW imaging algorithm is improved for planar antenna array,which is used to deal with approximate processing of imaging data under near-field condition.According to this modified algorithm,traditional matched filtering function and interpolation ha ve been replaced by proximity focusing and consistent focus function.It not only ensures the imaging quality but also improves the imaging efficiency.Secondly,many problems,such as a long time for scanning and high hardware costing,are caused by a lot of antenna array unit for the actual security imaging syste m.This motivates us to improve two reconstruction algorithm and models that based on edge detection.The algorithm is based on the framework of compressed sensing,which can recover the original image with a large probability.The framework uses non-coherent sampling to get a small amount of echo data which possess sparse priori characteristics.It is very significant to optimize the millimeter wave security imaging system.Then,the two improved algorithms are as follows:(1)The edge-CS algorithm based on edge guidance is improved for three-dimensional millimeter wave echo data,which satisfies the high requirement of security check for target detail feature detection.3D Edge-CS imaging algorithm model is based on the three dimensional total variation CS reconstruction(3D CS-TV)algorithm,which can acquire the better edge detail than 3D CS-TV algorithm reconstruction results.In the case of a reconstructed image with a sampling rate of 20%,the error is reduced by 23.12%.(2)Since the used gradient operator is very sensitive to noise,it is lead to the problem of inaccurate edge information.In this thesis,an edge nested weighting sparse imaging algorithm model is proposed based on 3D Edge-CS algorithm,which is based on the acquisition of edge information and then combined wit h the edge of the support of the weight of learning.In the case of a reconstructed image with a sampling rate of 20%,the error is reduced by 13.64%.Lastly,in this thesis,the improved algorithms are used to solve the problem of slow imaging speed and high hardware cost.By massive simulation and experimental experiments,it is proved that the improved algorithm has a considerable advantage compared with other traditional algorithms to some extent.The algorithm provides a new idea for the practical application of the follow-up active millimeter wave security imaging.