基于组稀疏压缩感知的穿墙雷达成像研究

发布时间：2018-03-07 09:45

本文选题：穿墙雷达成像　切入点：压缩感知　出处：《南京理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：穿墙雷达能够对隐藏在墙体或建筑物后的目标进行探测和成像,在灾后救援、反恐作战等诸多领域都具有广泛的应用。穿墙雷达成像作为一种新的复杂环境下的雷达成像技术,在诸多方面仍面临挑战。一方面,高分辨成像需求增加了系统复杂度(如大带宽、多天线等),而复杂的系统导致大量的测量数据,给系统的存储、传输、处理等环节造成了很大负担。另一方面,由于墙体对电磁波的反射作用,接收机会收到来自多个传播路径的目标回波,由此形成的多径效应会在重建图像中产生不期望的虚像,从而影响成像质量。针对上述问题,本文主要研究基于压缩感知理论的穿墙雷达高分辨成像技术和多径抑制技术。本文在利用穿墙雷达接收信号所呈现的组稀疏性基础上,主要开展基于组稀疏性的穿墙雷达成像技术研究。为有效利用穿墙雷达的组稀疏性,本文采用贝叶斯压缩感知实现了低数据量下的高分辨成像。穿墙雷达的多径效应同样呈现组稀疏特性,本文提出了一种基于组稀疏的多径分集穿墙雷达成像方法。本论文主要工作如下:(1)简述基于组稀疏压缩感知的穿墙雷达成像理论框架。本文在介绍穿墙雷达成像的基本原理和压缩感知理论基础上,针对穿墙雷达接收信号模型,根据其具有组稀疏性的特点,研究了基于组稀疏性的穿墙雷达成像方法,并利用两种组稀疏重构算法,可分离逼近稀疏重构法(SpaRSA)和块正交匹配追踪法(BOMP),验证了基于组稀疏性的穿墙雷达成像性能。(2)发展基于贝叶斯压缩感知的组稀疏穿墙雷达成像技术。为有效利用穿墙雷达成像的组稀疏性,本文将贝叶斯压缩感知理论应用于组稀疏穿墙雷达成像,在构建组稀疏贝叶斯模型基础上,采用复多任务贝叶斯压缩感知方法(CMT-BCS)迭代更新贝叶斯模型参数和重构图像,从而实现组稀疏穿墙雷达成像。仿真结果表明,基于贝叶斯压缩感知的穿墙雷达成像能够准确地对墙后场景图像进行重构,并且计算效率较高。(3)提出基于组稀疏贝叶斯压缩感知的多径分集穿墙雷达成像技术。针对穿墙雷达多径效应问题,本文在深入研究穿墙雷达多径传播模式基础上,揭示其同样具有组稀疏性,因此有效利用多径信号的组稀疏性,不仅可抑制多径所产生的虚像,还可提高穿墙雷达的成像性能。仿真实验结果表明基于组稀疏压缩感知的多径分集穿墙雷达成像在穿墙多径环境下仍能够准确的重构场景中的目标,同时有效地抑制了虚像,取得了高质量的成像结果。
[Abstract]:Wall-penetrating radar can detect and image targets hidden behind walls or buildings. It has been widely used in many fields, such as post-disaster rescue, anti-terrorist warfare, etc. As a new radar imaging technology in complex environment, penetrating wall radar imaging is a kind of radar imaging technology. On the one hand, high resolution imaging requirements increase system complexity (such as large bandwidth, multiple antennas, etc.), while complex systems lead to a large amount of measurement data, storage and transmission to the system. On the other hand, because of the reflection of the wall to the electromagnetic wave, the receiver will receive the target echo from multiple propagation paths, and the resulting multipath effect will produce an unwanted virtual image in the reconstructed image. In order to affect the imaging quality, this paper mainly studies the high-resolution imaging technology and multi-path suppression technology based on the compression sensing theory. In this paper, based on the group sparsity of the received signal from the penetrating radar, we mainly study the high resolution imaging technology and the multi-path suppression technology based on the compression perception theory. In order to make effective use of the thinness of penetrating wall radar, the imaging technology of penetrating wall radar based on group sparsity is studied. In this paper, Bayesian compression sensing is used to realize high resolution imaging under low data volume. The multipath effect of wall-penetrating radar is also sparse. In this paper, we propose a multipath diversity penetrating wall radar imaging method based on sparse group. The main work of this thesis is as follows: 1) briefly introduce the theoretical framework of penetrating wall radar imaging based on sparse compression sensing. In this paper, we introduce the imaging of penetrating wall radar. On the basis of the theory of compression perception, Aiming at the receiving signal model of penetrating radar, according to its characteristics of group sparsity, the imaging method of penetrating radar based on group sparsity is studied, and two sparse reconstruction algorithms are used. Separable approximation sparse reconstruction method (SpaRSA) and block orthogonal matching tracking method (BOMPA) are used to verify the imaging performance of wall penetrating radar based on group sparsity. The development of group sparse penetrating radar imaging technology based on Bayesian compression sensing is presented. Sparse group of wall radar imaging, In this paper, Bayesian compression sensing theory is applied to group sparse wall radar imaging. On the basis of constructing group sparse Bayesian model, complex multitask Bayesian compression sensing method (CMT-BCSS) is used to iteratively update Bayesian model parameters and reconstruct images. The simulation results show that the radar imaging based on Bayesian compression perception can accurately reconstruct the image of the back-wall scene. Furthermore, a new multi-path diversity penetrating radar imaging technique based on sparse Bayesian compression sensing is proposed. Aiming at the multipath effect of the penetrating radar, the multipath propagation mode of the penetrating radar is studied in this paper. It is revealed that the multipath signal is also group sparse. Therefore, using the group sparsity of multipath signal effectively can not only suppress the virtual image produced by multipath. The simulation results show that the multipath diversity penetrating radar imaging based on sparse compression sensing can reconstruct the target of the scene accurately in the multi-path environment, and effectively suppress the virtual image at the same time. High quality imaging results are obtained.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN957.52

【参考文献】