基于压缩感知的多维度雷达成像方法研究
发布时间：2018-10-29 13:14
【摘要】：成像雷达能根据目标的电磁散射回波反演目标的散射率分布,对于目标识别等应用具有重要意义。为了更全面精细地刻画雷达目标散射特性,多维度雷达成像方法,包括被动双站雷达成像、高分辨全极化雷达成像以及三维雷达成像应运而生。新兴的压缩感知理论提供了全新的信号采集框架,能从欠采样数据中精确重构原信号。因此,将压缩感知理论应用于雷达成像,有望解决信号采样率高、数据量大以及不完整采样下成像等雷达所面临的问题。本文以提高雷达成像能力为目的,以基于压缩感知理论的雷达成像方法研究为主线,深入研究了被动双站ISAR成像、高分辨全极化ISAR成像以及三维雷达成像的理论和方法。第一章阐述了论文的研究背景及意义,总结和归纳了压缩感知理论以及压缩感知雷达成像方法的研究现状,在此基础上指出了多维度雷达成像所需解决的问题,最后介绍了本文的主要工作和内容安排。第二章描述了基于压缩感知的ISAR成像基本原理。首先对压缩感知理论的数学模型进行了简要回顾,其次采用频率步进信号推导了雷达目标二维ISAR成像的回波模型,然后将压缩感知方法应用于ISAR成像,探讨了两种稀疏采样方案,并选用合适的重构算法以及字典加密倍数,最后利用稀疏性先验实现高分辨图像重构,仿真数据和实测数据结果验证了压缩感知ISAR成像方法的有效性,为后续章节的研究提供了理论和方法基础。第三章针对被动双站ISAR成像中的栅瓣问题和图像分辨率不高的不足,研究了基于压缩感知的被动双站ISAR成像方法。首先基于Batches算法思想推导了基于机会照明源的被动双站ISAR成像信号模型,指出从统计平均的角度看,最终的成像模型同样能表达成适于压缩感知处理的二维矩阵形式;在目标稀疏性的先验信息约束下,通过求解最优化问题得到目标的散射中心参数,进而计算完成缺失频带补偿的完整数据,然后利用距离-多普勒成像算法获得最终的ISAR像。进一步地,在获得的散射中心参数基础上,根据信号模型人为地外推测量数据,使得ISAR成像的分辨率得到增强。仿真数据和实测数据处理结果表明该方法能有效降低栅瓣的影响,并且能够提高成像结果的分辨率。第四章针对传统单极化处理方法不能保证散射中心在不同极化下数量、位置的一致性,研究了基于压缩感知的高分辨全极化ISAR成像方法。首先指出全极化下的ISAR像具有相同的稀疏性支撑,也就是联合稀疏性,从而将全极化ISAR成像问题转化为二维多测量稀疏恢复问题;然后为了表征这种联合稀疏性,定义了两类混合范数,并且用连续的高斯函数近似该混合范数;最后求解由混合范数约束的最优化问题,得到ISAR成像结果。仿真数据、暗室实测数据及外场实测数据的实验结果表明,两种基于混合范数优化的全极化ISAR成像方法不仅能利用欠采样数据获得高分辨ISAR像,而且成像结果中散射中心是对齐的,有利于后续的目标识别等应用。第五章研究了基于压缩感知的三维雷达成像方法,包括干涉ISAR成像和转台三维ISAR成像两部分。在干涉ISAR成像中,借鉴第四章的研究思路,认为基线对应的两幅ISAR像有相同的稀疏性支撑,进而定义两类全局稀疏性,并以此为约束求解优化问题,获得高质量的ISAR像,进一步作干涉处理即可得到目标的三维重构结果。对于三维转台ISAR成像,针对传统向量化压缩感知计算复杂度高、内存消耗大的不足,利用三维数据的结构特性提出了降维压缩感知方法和张量压缩感知方法,前者将三维观测数据展开为矩阵形式处理而后者直接对三维数据进行处理,两种方法均大大提高了算法的运算效率,并能够明显降低内存消耗,有望将其应用到大尺寸目标成像中。点目标仿真实验和电磁软件计算数据结果表明所提方法能有效获得目标三维像。最后在第六章对全文进行了总结,并展望了下一步的研究工作。
[Abstract]:The imaging radar can obtain the scattering rate distribution of the target according to the target electromagnetic scattering echo, and has important significance for the target recognition and the like. In order to depict radar target scattering characteristic more fully, multi-dimension radar imaging method includes passive double-station radar imaging, high-resolution full-polarization radar imaging and three-dimensional radar imaging. The new compression-sensing theory provides a new signal acquisition framework, which can reconstruct the original signal accurately from under-sampled data. Therefore, the compression-sensing theory is applied to radar imaging, which is expected to solve the problems faced by radar such as high sampling rate, large data volume and incomplete sampling. In order to improve radar imaging capability, this paper studies the theory and method of passive bistatic ISAR imaging, high resolution full polarization ISAR imaging and three-dimensional radar imaging based on the research of radar imaging method based on the theory of compression sensing. The first chapter expounds the background and significance of the thesis, sums up and summarizes the research status of the compression sensing theory and the compression sensing radar imaging method, and then points out the problems needed to solve the multi-dimensional radar imaging, and finally introduces the main work and the content arrangement in this paper. The second chapter describes the basic principle of ISAR imaging based on compression sensing. Firstly, the mathematical model of the compression sensing theory is briefly reviewed, and then the echo model of the radar target two-dimensional ISAR imaging is derived by using the frequency step signal, and then the compression sensing method is applied to ISAR imaging, and two sparse sampling schemes are discussed. In this paper, a suitable reconstruction algorithm and a dictionary encryption multiple are selected, and the validity of the compression-sensing ISAR imaging method is verified by using the sparse priori to realize the reconstruction of the high-resolution image, the simulation data and the measured data. The theoretical and methodological bases are provided for the research of the following chapters. In chapter three, the problem of grating lobe and image resolution in passive bistatic ISAR imaging are studied, and a passive bistatic ISAR imaging method based on compression sensing is studied. Firstly, based on the Batcheles algorithm, a passive bistatic ISAR imaging signal model based on the opportunity illumination source is derived. It is pointed out that the final imaging model can be expressed in a two-dimensional matrix form suitable for the compression sensing process from the statistical average point of view; under the prior information constraint of the target sparsity, By solving the optimization problem, the scattering center parameters of the target are obtained, and then the complete data for completing the missing band compensation is calculated, and then the final ISAR image is obtained by using the distance-Doppler imaging algorithm. Further, based on the obtained scattering center parameters, the measurement data is artificially extrapolated according to the signal model, so that the resolution of ISAR imaging is enhanced. The simulation data and the data processing results show that the method can effectively reduce the influence of the grating lobes, and can improve the resolution of the imaging results. In chapter 4, aiming at the consistency of the number and position of scattering center under different polarization, the method of high resolution full polarization ISAR imaging based on compression sensing is studied. First, it is pointed out that the ISAR image under full polarization has the same sparsity support, that is, the joint sparsity, so that the whole-polarized ISAR imaging problem is converted into two-dimensional multi-measurement sparse recovery problem; then, in order to characterize this joint sparsity, two kinds of mixed norms are defined. and solving the optimization problem constrained by the mixed norm to obtain ISAR imaging results. The experimental results of simulation data, darkroom test data and field measured data show that the two-polarization ISAR imaging method based on hybrid norm optimization can not only obtain the high resolution ISAR image by using the undersampled data, but also the scattering centers in the imaging results are aligned. and is favorable for subsequent target recognition and the like. In chapter five, a three-dimensional radar imaging method based on compression sensing is studied, including interference ISAR imaging and turntable three-dimensional ISAR imaging. In the ISAR imaging, the two ISAR images corresponding to the baseline are considered to have the same sparsity support, and then two kinds of global sparsity are defined, and then the optimization problem is solved by taking the constraint solving optimization problem to obtain the high-quality ISAR image. and further performing interference processing to obtain a three-dimensional reconstruction result of the target. For the three-dimensional turntable ISAR imaging, aiming at the defects of high computational complexity and large memory consumption to the traditional vectorized compression sensing calculation, a method for reducing dimension compression sensing and a tensor compression sensing method are proposed by utilizing the structural characteristics of the three-dimensional data. In the former, the three-dimensional observation data is expanded into matrix form processing and the latter directly processes the three-dimensional data. Both methods greatly improve the operation efficiency of the algorithm, and can obviously reduce the memory consumption and can be applied to large-scale target imaging. The simulation experiment of point target and the calculation data of electromagnetic software show that the proposed method can get the target three-dimensional image effectively. Finally, the paper summarizes the whole thesis in Chapter 6, and looks forward to the next research work.
【学位授予单位】：国防科学技术大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN957.52
，
本文编号：2297791
[Abstract]:The imaging radar can obtain the scattering rate distribution of the target according to the target electromagnetic scattering echo, and has important significance for the target recognition and the like. In order to depict radar target scattering characteristic more fully, multi-dimension radar imaging method includes passive double-station radar imaging, high-resolution full-polarization radar imaging and three-dimensional radar imaging. The new compression-sensing theory provides a new signal acquisition framework, which can reconstruct the original signal accurately from under-sampled data. Therefore, the compression-sensing theory is applied to radar imaging, which is expected to solve the problems faced by radar such as high sampling rate, large data volume and incomplete sampling. In order to improve radar imaging capability, this paper studies the theory and method of passive bistatic ISAR imaging, high resolution full polarization ISAR imaging and three-dimensional radar imaging based on the research of radar imaging method based on the theory of compression sensing. The first chapter expounds the background and significance of the thesis, sums up and summarizes the research status of the compression sensing theory and the compression sensing radar imaging method, and then points out the problems needed to solve the multi-dimensional radar imaging, and finally introduces the main work and the content arrangement in this paper. The second chapter describes the basic principle of ISAR imaging based on compression sensing. Firstly, the mathematical model of the compression sensing theory is briefly reviewed, and then the echo model of the radar target two-dimensional ISAR imaging is derived by using the frequency step signal, and then the compression sensing method is applied to ISAR imaging, and two sparse sampling schemes are discussed. In this paper, a suitable reconstruction algorithm and a dictionary encryption multiple are selected, and the validity of the compression-sensing ISAR imaging method is verified by using the sparse priori to realize the reconstruction of the high-resolution image, the simulation data and the measured data. The theoretical and methodological bases are provided for the research of the following chapters. In chapter three, the problem of grating lobe and image resolution in passive bistatic ISAR imaging are studied, and a passive bistatic ISAR imaging method based on compression sensing is studied. Firstly, based on the Batcheles algorithm, a passive bistatic ISAR imaging signal model based on the opportunity illumination source is derived. It is pointed out that the final imaging model can be expressed in a two-dimensional matrix form suitable for the compression sensing process from the statistical average point of view; under the prior information constraint of the target sparsity, By solving the optimization problem, the scattering center parameters of the target are obtained, and then the complete data for completing the missing band compensation is calculated, and then the final ISAR image is obtained by using the distance-Doppler imaging algorithm. Further, based on the obtained scattering center parameters, the measurement data is artificially extrapolated according to the signal model, so that the resolution of ISAR imaging is enhanced. The simulation data and the data processing results show that the method can effectively reduce the influence of the grating lobes, and can improve the resolution of the imaging results. In chapter 4, aiming at the consistency of the number and position of scattering center under different polarization, the method of high resolution full polarization ISAR imaging based on compression sensing is studied. First, it is pointed out that the ISAR image under full polarization has the same sparsity support, that is, the joint sparsity, so that the whole-polarized ISAR imaging problem is converted into two-dimensional multi-measurement sparse recovery problem; then, in order to characterize this joint sparsity, two kinds of mixed norms are defined. and solving the optimization problem constrained by the mixed norm to obtain ISAR imaging results. The experimental results of simulation data, darkroom test data and field measured data show that the two-polarization ISAR imaging method based on hybrid norm optimization can not only obtain the high resolution ISAR image by using the undersampled data, but also the scattering centers in the imaging results are aligned. and is favorable for subsequent target recognition and the like. In chapter five, a three-dimensional radar imaging method based on compression sensing is studied, including interference ISAR imaging and turntable three-dimensional ISAR imaging. In the ISAR imaging, the two ISAR images corresponding to the baseline are considered to have the same sparsity support, and then two kinds of global sparsity are defined, and then the optimization problem is solved by taking the constraint solving optimization problem to obtain the high-quality ISAR image. and further performing interference processing to obtain a three-dimensional reconstruction result of the target. For the three-dimensional turntable ISAR imaging, aiming at the defects of high computational complexity and large memory consumption to the traditional vectorized compression sensing calculation, a method for reducing dimension compression sensing and a tensor compression sensing method are proposed by utilizing the structural characteristics of the three-dimensional data. In the former, the three-dimensional observation data is expanded into matrix form processing and the latter directly processes the three-dimensional data. Both methods greatly improve the operation efficiency of the algorithm, and can obviously reduce the memory consumption and can be applied to large-scale target imaging. The simulation experiment of point target and the calculation data of electromagnetic software show that the proposed method can get the target three-dimensional image effectively. Finally, the paper summarizes the whole thesis in Chapter 6, and looks forward to the next research work.
【学位授予单位】：国防科学技术大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN957.52
，
本文编号：2297791
本文链接：https://www.wllwen.com/kejilunwen/wltx/2297791.html
