Near-field Millimeter Wave SAR Imaging System Design And Algorithm Research

Synthetic Aperture Radar(SAR)is an imaging radar with high resolution,which obtains the effect of equivalent large aperture radar resolution by forming a radar virtual antenna through relative motion with the target.With the continuous development of SAR technology and its increasing resolution,millimetre wave SAR with these advantages is widely used in near field applications such as security screening and assisted driving.Therefore,the study of near-field millimetre wave SAR imaging is of great importance for practical applications.Since conventional SAR imaging platforms are expensive and large in size and not easily movable,a low-cost,compact near-field SAR imaging system is built based on an integrated chip-based millimetre wave radar and used as an experimental platform for millimetre wave radar imaging to carry out research on SAR imaging-related algorithms.To address the problems of long computing time and high system complexity of existing imaging algorithms,a fast near-field 3D SAR imaging algorithm is proposed,and a neural network combined with data enhancement is used for SAR image target recognition.Details of the research are as follows.(1)A millimeter-wave SAR imaging system is built based on a millimeter-wave radar,which can achieve two-dimensional plane scanning.At the same time,a supporting visualization operation software is developed.This software can configure the parameters of the radar guide rail,including the direction of travel,speed and distance travelled,and display the imaging results of the target.The platform collects data on the target and combines it with the Back Projection(BP)imaging algorithm to successfully obtain the twodimensional image result of the target,verifying the feasibility of this platform.(2)This paper proposes a fast near-field 3D SAR imaging algorithm to address the problems of mixing,contour blur,and excessively long imaging algorithm time for 2D imaging.First,the algorithm establishes a distance unit division model,uses synthetic aperture technology in the directional and height dimensions to obtain high-resolution 2D images at different distance units,and then adds the 2D images to obtain the 3D imaging result of the target.At the same time,the geometric model of the algorithm is established,and the theoretical model of the algorithm is derived.(3)Based on the theoretical model of the imaging algorithm,the radar waveform parameters are set,and simulations and real experiments are conducted.The imaging results are compared with those of other imaging algorithms,and the results demonstrate that the algorithm proposed in this paper effectively reconstructs the target,reduces operation time,and ensures high imaging quality.To address the problem of small sample size of SAR images,this paper adopts a data enhancement approach to expand the sample size of SAR images and combines neural networks to recognise their targets.The feasibility of the recognition algorithm is also verified using both publicly available and measured datasets.