多通道SAR成像理论与方法研究
发布时间：2018-11-19 09:25
【摘要】：近年来,以多极化/多天线为代表的多通道SAR成为了SAR发展的一大趋势。多极化信息的获取使SAR对目标散射特性的描述更加完整;多天线使SAR可以从多个不同的角度对场景进行观测,扩大了SAR的信息来源,拓展了SAR的应用范围。论文立足于多通道SAR体制,以提高SAR的战场情报获取能力和处理能力为目标,深入开展了多通道SAR成像理论与方法研究,其研究成果有助于拓展SAR在超分辨成像、前视成像、动目标成像等领域的应用。第一章阐述了课题研究的背景和意义,归纳了极化SAR相干斑滤波、极化SAR超分辨成像、前视SAR成像、多通道SAR-GMTI理论和系统的研究现状,并分析了当前研究所存在的问题,最后概括了论文的主要工作。第二章研究了基于均值移动的极化SAR相干斑抑制方法。首先回顾了相干斑的形成机理、统计特性和几种经典的极化SAR相干斑滤波算法。接下来,针对当前极化SAR相干斑滤波算法不能很好的保护图像细节信息和极化散射特性的不足,提出了一种基于均值移动的极化SAR相干斑滤波算法——PDAN算法。该算法采用了空间域和值域的联合滤波,充分利用了极化信息和幅度信息,满足了极化SAR相干斑滤波的基本原则。基于德国DLR的L波段E-SAR数据和中国电子科技集团38所的X波段机载极化SAR数据的实验结果表明:PDAN算法比当前广泛使用的精致Lee极化滤波算法和IDAN滤波算法具有更平滑的相干斑滤波效果以及更好的图像细节、边界信息、极化散射特性保持能力。第三章围绕着“改善抗噪性能”和“超分辨成像”两个问题开展了基于稀疏重建的极化SAR超分辨成像方法研究。首先,提出了一种新的SAR成像处理框架,该处理框架可以有效克服杂波和噪声的影响。在此基础之上,基于压缩感知理论和非二次最优化技术提出了一种迭代超分辨成像算法——IFCS算法。基于仿真数据的成像结果表明,IFCS算法在抗噪性能和超分辨性能方面都优于现有算法。设计并构建了X波段步进频轨道SAR实验系统,开展了成像实验,实验场景中包含有角反射器和车辆目标。实验结果表明,对于结构简单的角反射器目标,IFCS算法能够得到背景干净,旁瓣很低的成像结果。对于具有复杂结构的车辆目标,IFCS算法在克服旁瓣影响的同时获得了更多的目标细节信息。第四章深入开展了高速多通道前视SAR成像算法研究。首先,针对当前多通道前视SAR成像算法只适用于低速平台的问题,建立了高速多通道前视SAR信号模型,并以此为依据,提出了适用于高速多通道前视SAR的DBS和RD成像算法,解决了高速运动平台的运动补偿问题。接下来,为了解决多通道前视SAR方位向分辨率受到平台物理尺寸制约的问题,将压缩感知理论创新的应用在多通道前视SAR成像领域,提出了两种适用于高速多通道前视SAR的超分辨成像算法。其中,二维压缩感知超分辨成像算法具有不低于5倍的超分辨能力,但是需要对场景做二维剖分,计算量较大。方位向压缩感知超分辨成像算法虽然成像精度和分辨率略有降低,但大大降低了运算量,适合于较大场景的成像。基于Mini SAR实测数据的分布式目标成像仿真表明:本算法对建筑物目标和飞机目标均具有很好的重构性能。第五章结合国内首部7通道SAR-GMTI系统的数据处理工作,深入开展了多通道SAR-GMTI处理方法研究。首先建立了运动目标在SAR系统中的回波模型,并利用理论推导和计算机仿真的手段对运动目标的成像效果进行了分析。在此基础上,进一步分析了运动目标在不同通道中的成像差异,建立了多通道SAR-GMTI信号模型。接下来,在回顾DPCA、ATI、STAP这3种典型的多通道SAR-GMTI处理算法的基础上,开展了7通道SAR-GMTI系统实测数据的处理,重点研究了惯导信息缺失情况下的自聚焦、通道均衡以及降维STAP处理方法。提出了一种实用的通道均衡方法——线性均衡方法,取得了不错的通道均衡效果。最终,完成了国内首部7通道SAR-GMTI系统的数据处理工作,得到了很好的处理结果。第六章总结了全文的主要工作和创新点,并对需要进一步研究的问题进行了展望。
[Abstract]:In recent years, the multi-channel SAR represented by multi-polarization/ multi-antenna has become a major trend of the development of SAR. The acquisition of multi-polarization information makes the description of the SAR to the target scattering characteristic more complete; the multi-antenna enables the SAR to observe the scene from a plurality of different angles, enlarges the information source of the SAR, and expands the application range of the SAR. In this paper, based on the multi-channel SAR system, the research on the multi-channel SAR imaging theory and method is carried out in order to improve the information acquisition ability and processing ability of the battlefield information of the SAR, and the research results can help to expand the application of the SAR in the fields of super-resolution imaging, front-view imaging, moving target imaging and the like. In the first chapter, the background and significance of the subject research are described, and the research status of the polarization SAR coherent spot filter, the polarization SAR super-resolution imaging, the front-view SAR imaging, the multi-channel SAR-GMTI theory and the system are summarized, and the problems existing in the current research institute are analyzed, and the main work of the paper is finally summarized. In the second chapter, the method of polarization SAR coherent spot suppression based on mean-shift is studied. First, the formation mechanism, statistical property and several classical polarization SAR coherent spot filtering algorithms of coherent spot are reviewed. Next, aiming at the insufficient protection of the image detail information and the polarization scattering property of the current polarized SAR coherent spot filtering algorithm, a polarized SAR coherent speckle filtering algorithm based on the mean movement is proposed. In this paper, the combined filtering of the spatial domain and the range is used, the polarization information and the amplitude information are fully utilized, and the basic principle of the coherent speckle filtering of the polarized SAR is met. The experimental results of the L-band E-SAR data based on DLR and the X-band on-board polarized SAR data of China E & T Group 38 show that: The PDAN algorithm has a smoother coherence-spot filtering effect and better image detail, boundary information, and polarization-scattering property retention capability than the sophisticated Lee-polarized filter algorithm and the IDAN filtering algorithm which are widely used in the present invention. In the third chapter, based on the two problems of 鈥渋mprove anti-noise performance鈥，
本文编号：2341865
[Abstract]:In recent years, the multi-channel SAR represented by multi-polarization/ multi-antenna has become a major trend of the development of SAR. The acquisition of multi-polarization information makes the description of the SAR to the target scattering characteristic more complete; the multi-antenna enables the SAR to observe the scene from a plurality of different angles, enlarges the information source of the SAR, and expands the application range of the SAR. In this paper, based on the multi-channel SAR system, the research on the multi-channel SAR imaging theory and method is carried out in order to improve the information acquisition ability and processing ability of the battlefield information of the SAR, and the research results can help to expand the application of the SAR in the fields of super-resolution imaging, front-view imaging, moving target imaging and the like. In the first chapter, the background and significance of the subject research are described, and the research status of the polarization SAR coherent spot filter, the polarization SAR super-resolution imaging, the front-view SAR imaging, the multi-channel SAR-GMTI theory and the system are summarized, and the problems existing in the current research institute are analyzed, and the main work of the paper is finally summarized. In the second chapter, the method of polarization SAR coherent spot suppression based on mean-shift is studied. First, the formation mechanism, statistical property and several classical polarization SAR coherent spot filtering algorithms of coherent spot are reviewed. Next, aiming at the insufficient protection of the image detail information and the polarization scattering property of the current polarized SAR coherent spot filtering algorithm, a polarized SAR coherent speckle filtering algorithm based on the mean movement is proposed. In this paper, the combined filtering of the spatial domain and the range is used, the polarization information and the amplitude information are fully utilized, and the basic principle of the coherent speckle filtering of the polarized SAR is met. The experimental results of the L-band E-SAR data based on DLR and the X-band on-board polarized SAR data of China E & T Group 38 show that: The PDAN algorithm has a smoother coherence-spot filtering effect and better image detail, boundary information, and polarization-scattering property retention capability than the sophisticated Lee-polarized filter algorithm and the IDAN filtering algorithm which are widely used in the present invention. In the third chapter, based on the two problems of 鈥渋mprove anti-noise performance鈥，
本文编号：2341865
本文链接：https://www.wllwen.com/kejilunwen/wltx/2341865.html
