【作者】 安道祥；

【导师】 周智敏； 黄晓涛；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2011， 博士

【摘要】 获得高质量和高分辨率图像是合成孔径雷达(Synthetic Aperture Radar,SAR)追求的目标。随着SAR技术应用领域的推广,要求SAR系统可在不同体制和工作模式下具有高分辨率成像能力,并能在各种实际飞行条件下获得令人满意的高质量图像。本文结合工程实际应用,研究了低频超宽带SAR(Ultra-Wideband SAR,UWB SAR)、聚束SAR和大斜视SAR三种工作体制在高分辨率情况下的若干理论和技术问题。本文主要内容概括如下:1.统一了不同SAR频域成像算法的理论推导,分析了不同频域算法对高分辨率SAR的成像性能。针对SAR频域成像算法种类繁多,推导形式不统一的问题,本文基于SAR通用回波信号模型,给出了不同频域算法的统一形式推导。基于该推导过程,从SAR信号处理角度重新解释了扩展Omega-K算法(Extended Omega-K Algorithm,EOKA)的两维分离聚焦成像原理,较之前解释方法更容易理解。最后,以低频UWB SAR为例,对比分析了不同频域算法对高分辨率SAR的成像性能,所得结论为后续研究奠定了基础。2.从理论上分析了平台运动误差对SAR成像处理的影响。具体工作为:(1)从理论上分析了航向速度误差对SAR成像处理的影响,得出随着SAR分辨率的提高或工作波段的降低,航向速度误差的影响越来越大的结论。(2)基于正弦平动误差模型,推导了非理想情况下的回波信号频谱形式,指出了不同平动误差造成SAR图像出现不同散焦现象的根本原因,并从理论上证明平动误差的幅度或频率越大,对SAR成像处理的影响越严重。所得分析结论为后续研究奠定了基础。3.研究了基于回波数据的高分辨率UWB SAR运动补偿方法。针对无运动测量数据情况下的UWB SAR运动补偿问题,提出了基于回波数据的运动补偿方法。具体工作为:(1)根据UWB SAR实测数据特点,改进了传统基于多普勒调频率估计的运动补偿方法,提高了对UWB SAR实测数据的补偿效果。(2)针对平台速度误差变化导致回波距离聚焦精度下降的问题,提出了子孔径修正Stolt插值法。(3)为补偿UWB SAR图像中的二维空变高次相位误差,提出了图像分块自聚焦处理法,提高了UWB SAR实测图像的整体聚焦质量。4.研究了基于运动测量数据和回波数据的小型机载高分辨率UWB SAR运动补偿方法。针对小型机载UWB SAR运动误差频率高,幅度大,不易补偿的问题,提出了结合低精度全球定位系统(Global Position System,GPS)测量数据和回波数据的三级运动补偿法。为提高基于GPS数据的粗补偿效果,提出了改进两步运动补偿法。该方法在保持对中/低频运动误差良好补偿效果的同时,具有更好的高频运动误差补偿性能。通过采用基于低通滤波器的位移数据平滑处理,消除了GPS系统测量误差的影响,并给出了低通滤波器截止频率的确定方法。5.研究了基于去调频技术的高分辨率聚束SAR成像方法。为解决具有方位谱混叠现象的斜视聚束SAR成像问题,提出一种扩展两步式成像法。具体工作为:(1)深入分析了斜视聚束SAR的方位谱混叠现象,找出了斜视角影响方位粗聚焦的根本原因。(2)提出了多普勒中心非线性平移校正法,消除了斜视角的影响,实现了斜视聚束SAR的扩展两步式成像处理。(3)为解决粗聚焦回波的方位滤波失配问题,提出了方位子带精聚焦法,提高了扩展两步式成像法对大场景高分辨率斜视聚束SAR的成像性能。6.研究了高分辨率大斜视条带SAR成像方法。为解决宽测绘带高分辨率斜视SAR的距离弯曲校正问题,提出了结合距离子带的一致距离弯曲校正法,提高了回波信号的距离聚焦精度,并给出了距离子带的确定方法。针对线性距离走动校正法中存在的场景聚焦深度限制问题,提出了基于非线性调频变标(NonLinear Chirp Scaling,NLCS)技术的方位滤波法,实现了对宽测绘带高分辨率大斜视SAR的高精度方位滤波处理。本文的部分研究成果已经应用于国内首部机载高分辨率低频UWB SAR系统和首部小型机载高分辨率低频UWB SAR系统的实测数据成像处理中,获得了大量高质量高分辨率的实测图像,从而证明了本文研究结果的良好实际应用价值。更多还原

【Abstract】 One of the goals of Synthetic Aperture Radar (SAR) is to obtain the high resolution image with well focused quality. As the wide application of SAR techniques, all kinds of SAR systems are required to have high resolution imaging ability in different operation mode, and obtain the well focused images in various flying conditions in practice. In this dissertation, some theoretical and technical problems are studied, which are met in the application of three different SAR systems when they working on the high resolution imaging mode. The three SAR systems include ultra-wideband SAR (UWB SAR), spotlight SAR and highly squinted SAR.The main content of this dissertation is summarized as follows:1. Generalized expression of SAR frequency domain algorithms is derived and the imaging performances of different algorithms for the high resolution SAR imaging are analyzed. Based on the SAR universal echo signal model, the generalized deduction of SAR frequency algorithms is presented. With the deduction, a new explanation of SAR imaging principle of the extended Omega-K algorithm (EOKA) from signal processing viewpoint is given, which is easier to be understood as comprared to the former explanations. At last, using the example of UWB SAR, comparison analyses of the performances of different frequency algorithms for high resolution SAR imaging are carried out, which form the basis for following study.2. Analyses of radar platform motion error on SAR imaging processing are carried out. The main work includes: (1) Theory analyses of the forward velocity error on SAR imaging are carried out. Based on the results, we conclude that as the SAR resolution higher or carrier frequency lower, the influence of forward velocity error worse. (2) Based on the sinusoidal trajectory deviation model, the expression of echo spectrum obtained in nonideal conditions is derived, and the season that why different trajectory deviations cause different defocus phenomenon in SAR image is found out. Moreover, a conclusion is drawn by thoery confirmation that as the amplitude or frequency of the trajectory deviation increases, it has worse influence on the SAR imaging. The analysis results form the basis for following study.3. The motion compensation (MoCo) method for high resolution UWB SAR based on raw data is studied. For the MoCo problem of low frequency UWB SAR without available measured motion data, a MoCo method based on raw data is proposed. The main works include: (1) According to the characteristic of UWB SAR raw data, a modified MoCo method using the Doppler frequency rate estimation is proposed, which improves the MoCo accuracy of UWB SAR raw data. (2) In order to remove the impacts of forward velocity error on echo range focusing, a subaperture modified Stolt interpolation processing is proposed. (3) In order to compensate the spatial variant high-order phase error in the UWB SAR real image, an image subpacthes focused method is proposed, which improves the focused quality of whole real image.4. The MoCo method for high resolution UWB SAR mounted on small size aircraft based on the measured motion data and raw data is studied. In order to compensate the large amplitude and high frequency motion error, a three-step MoCo method based on the low accuracy Global Position System (GPS) data and raw data is proposed. In order to obtain the better coarse focused result based on GPS data, a modified two-step MoCo approach is proposed, which not only has better performance on compensating the low/moderate frequency motion errors, but also has better performance on compensating the high frequency motion errors. By smoothing the trjactory deviation data with lowpass filter, the impacts of measure error induced by the GPS system is removed, and the way of the cutoff frequency determination is given.5. The imaging algorithm for high resolution squinted spotlight SAR based on the deramping-based techniques is studied. In order to resovle the imaging problem of squinted spotlight SAR in presence of the azimuth spectrum folding phenomenon, an extended two-step focusing approach is proposed. The main works includes: (1) the spectrum folding problem encountered in squinted spotlight SAR case is analyzed, and the reason of the squint angle has impact on the azimuth coarse focusing is found out. (2) To overcome this problem, a nonlinear shift method for Doppler centroid correction is proposed, and the tranditional two-step focusing approach is successfully extended for the squinted spotlight SAR imaging. (3)To resolve the azimuth mismatched filtering problem of the coarse focused echo sginal, an azimuth subscenes precise focusing method is proposed, which improves the performance of the extended two-step focusing approach for the wide swath high resolution squinted spotlight SAR imaging.6. The imaging algorithm for high resolution highly squinted SAR is studied. In order to resolve the range cell curvature correction (RCCC) problem, a bulk RCCC method with integrated the range blocks is proposed, which improves the accuracy of the range focusing. To overcome the azimuth depth of focus (ADOF) problem induced by the linear range walk correction, a modified azimuth mathed filtering based on the nonlinear chirp scaling (NLCS) technique is proposed, which achieves the high accuracy azimuth compression for the wide swath high resolution higly squinted SAR.Some research results achieved in this dissertation have been applied on the raw data imaging, which are collected by the first domestic airborne high resolution low frequency UWB SAR system and the first domestic high resolution low frequency UWB SAR system mounted on small size aircraft, and a large number of high resolution images with high focused quality are obtained. The imaging results prove that the obtained research results have well applicable value in practice.更多还原