【作者】 吴明宇；

【导师】 吴顺君；

【作者基本信息】 西安电子科技大学 ， 信号与信息处理， 2013， 博士

【摘要】 合成孔径雷达(SAR)是一种以合成孔径原理和脉冲压缩技术为理论基础，具有二维高分辨率的成像雷达。SAR属于主动式微波遥感设备，具有全天候、全天时和远距离成像的特点，大大提高了雷达信息获取能力，为目标检测创造了前所未有的机会，是现代雷达的一项突破性成就，在军事、国防以及民用领域都得到了非常广泛的应用。作为一种先进的成像系统，SAR能够从空中实时提供类似光学照片的二维地物地貌图像，并且具有良好的空间分辨能力和辐射测量保真度。SAR对地面运动目标的检测和成像研究是在合成孔径雷达成像技术的基础上发展起来的，SAR最初是针对地面静止场景成像的，在SAR图像中，地面运动目标图像会出现距离走动、方位位置偏移、散焦等现象，动目标往往只能模糊的叠加在静止场景的SAR图像上。在实际的应用中，特别是在军事的应用上，地面运动目标检测(GMTI)是SAR系统必须具备的一项基本功能，人们总是希望通过SAR图像来进行地面运动目标检测，并将其标注在场景的SAR图像上，因此诞生了SAR-GMTI技术。沿航迹向多通道SAR比单通道SAR能够提供更多的系统自由度，对于实现GMTI功能而言，雷达阵列沿航向直线分布为最优构型，可以获得最优的杂波抑制性能，能够在地杂波或者海杂波中清楚地区分出运动目标并且能够正确地估计出其运动参数，因而广泛地应用于战场感知和侦察中。目前世界上许多国家都在大力发展多通道地面运动目标检测和定位技术，积极研究先进的运动目标检测理论，并努力寻求各种高效、实用的运动目标检测方法。本文在深入分析国内外SAR-GMTI研究成果的基础上，对合成孔径雷达动目标检测和误差估计进行了较为详细的研究和分析。重点针对方位多通道正侧视雷达实现GMTI的功能，开展了通道相位误差估计，对于SAR成像中拼接成大图的子带误差估计，运动目标检测和定位，以及参数估计等关键技术研究。论文以方位多通道高分辨率宽测绘带SAR系统为对象，研究的主要内容可概括为以下四个部分：1.结合多通道地面运动目标检测GMTI的合成孔径雷达SAR系统在对大面积地物成像的同时可以完成运动目标检测与参数估计，此时要求各通道间的特性完全一致。然而在实际情况中，由于各种各样的原因，各通道间的特性不可能完全一致，这些非理想因素将严重影响运动目标检测效果。论文针对这一问题，提出两种通道相位误差估计方法。方法一利用杂波的信号特征向量与其导向矢量的线性关系直接进行通道误差估计。方法二通过对回波数据进行方位重采样，然后利用杂波信号特征向量张成的空间(即信号子空间)与真实导向矢量张成的空间相同的原理进行误差估计。实验证明，两种方法均能有效地进行通道相位误差估计，并且方法二具有更高的估计精度。2.合成孔径雷达系统通过发射线性调频信号获得距离高分辨率图像，随着分辨率要求的提高，发射信号的带宽也随之提高，发射大带宽信号常用的解决方案为发射若干个不同载频的子带信号，然后通过数字信号处理的方法将各个子带信号进行合成，从而得到宽带信号。在实际情况中，由于系统硬件等不可避免地存在各种误差，导致各子带信号的不一致，从而影响合成信号质量，因此必须对各子带信号误差进行估计并补偿。论文针对子带间的重叠频谱问题，提出一种子带误差估计方法，在利用步进线性调频信号合成大带宽信号时，对各子带间的不一致性进行补偿。该方法通过对相邻子带间的重叠谱相位差进行拟合得到子带幅相误差和延时误差，并对各子带回波进行加窗处理以消除因子带重叠引起的成对回波。3.同时获取高分辨率宽测绘带(HRWS)合成孔径雷达(SAR)图像并进行地面运动目标检测(GMTI)是合成孔径雷达成像的发展趋势。然而传统单通道星载SAR系统受最小天线面积的限制，为得到宽距离测绘带而不产生距离模糊，要求脉冲重复频率不能太高，而高方位分辨率又要求系统以高脉冲重复频率工作，因此难以同时获取方位高分辨和距离宽测绘带。论文针对这个问题，在存在多普勒模糊的情况下，详细讨论了如何利用多通道SAR系统进行运动目标检测，并提出一种基于空时谱的运动目标检测方法，该方法根据杂波和动目标的空时谱关系构造方位匹配函数，方位匹配滤波后，再利用波束形成技术进行杂波抑制，然后对运动目标的斜距历程进行拟合得到运动目标的运动方向。接着对杂波抑制后的数据进行聚焦处理，得到模糊的运动目标图像，并利用恒虚警检测技术检测得到所有的模糊运动目标。最后根据模糊图像的空间关系和估计出的运动方向检测出真实目标。4.同时获取方位高分辨和距离宽测绘带是合成孔径雷达的发展趋势，论文在实现SAR地面运动目标检测的基础上，提出一种地面运动目标检测的运动参数估计方法。该方法首先利用空时自适应处理方法进行杂波抑制，然后再对杂波抑制后的结果进行距离压缩得到运动目标的运动轨迹，从而得到目标的真实方位位置。接着进行聚焦处理，利用多普勒模糊关系从模糊的运动目标中判断出真实目标，获得运动目标因沿视线方向速度引起的方位位置偏移，从而得到运动目标沿视线方向的速度。更多还原

【Abstract】 Synthetic Aperture Radar (SAR) is a kind of2D high resolution imaging radar，which is based on the principle of synthetic aperture and pulse compression technologyas the theoretical basis. SAR is a great achievement of modern radar with thecharacteristics of all-weather, all day and night, and long range. Radar imagingtechnique can enhance information acquisition ability of SAR, and provides a hithertounknown opportunity for target indication. SAR therefore exhibits great value in allmilitary, defense and civilian applications. As an advanced imaging system, SAR canprovide the two-dimensional image of a scene with good spatial resolution and highradiation measurement precision. The study of GMTI is developed based on SARimaging technology. Since traditional SAR radars only simply take the―picture‖of astationary scene, however, moving targets will be displayed on the SAR image of astationary scene as blurred images. A moving target with a radial velocity will exhibitdisplaced from its true position on the SAR image, and the motion in the azimuthdirection will cause its image to be dispersed along azimuth. In practical applications,especially in military applications, it is always expected to achieve Ground MovingTarget Indication (GMTI) by using SAR sensors and to relocate ground moving targetson the SAR image of the scene. So GMTI is the basic function must have for SARsystem. The technique is called SAR-GMTI. Since along-track multichannel SAR canprovide more degrees of freedom than single channel one. For the realization of GMTIfunction, the radar array along-track liner distribution is the optimal configuration.Along-track multichannel SAR has the powerful ability to suppress the clutter, todistinguish moving targets from the clutter coming from land or sea and to accuratelyestimate the motion parameters. Thus it has been used extensively for air-to-groundsurveillance and reconnaissance.In recent years, many countries in the world have been making great efforts todevelop spaceborne/airborne Ground Moving Target Indication (GMTI) systems basedon multi-channel SAR, research on new SAR-GMTI theories and explore highlyefficient detecting and locating technology. Based on the extensive investigation of theSAR-GMTI developments over the world, this dissertation presents a detailed researchand analysis on SAR-GMTI algorithms and error estimation, with the emphasis onimplementing SAR-GMTI function for a side-looking multi-channel SAR, includingchannel phase errors estimation, subband error estimation for splicing into a largerimage of SAR imaging, GMTI and location, and parameters estimation. Based on the spaceborne multi-channel HRWS SAR systems as the object of study,the main content of this dissertation is summarized as follows.1. Combined with GMTI, multichannel SAR system can complete moving targetindication and parameters estimation at the same time for large area imaging. In themoving target indication, the required characteristics of each channel must becompletely consistent. However, in practical situations, characteristics of each channelcannot be completely consistent because of various reasons. These non ideal factors willseriously affect the moving target indication effect. In order to solve this problem, twomethods are proposed to estimate channel phase errors for multi-channel SAR andGMTI system. The first method is based on the linear relationship between the cluttereigenvectors and its steering vectors. For the second method, the received echoes areresampled in azimuth, and the channel errors are obtained based on the fact that thespace spanned by the signal eigenvectors referred to as the signal subspace is equal tothat by the practical steering vectors. The simulated and ground-based real datademonstrate the validity of the two methods, and show that the second method hasbetter performance.2. SAR system can obtain high range resolution image by transmitting linear FMsignal. As the resolution requirements increasing, the bandwidth of signal also increases.The common solution for transmitting a large bandwidth signal is to transmit severaldifferent carrier frequency subband signals, then compose each subband signals bymeans of digital signal processing to obtain the large bandwidth signal. In practicalsituations, each subband signal is not consistent due to the unavoidable existence ofvarious errors in system hardware, and then the result will influence the quality of thesynthesized signal. So it is necessary to estimate and compensate for each subbandsignals. In this paper, a novel method is proposed to estimate the sub-band errorsdepending on the overlap spectra between sub-band signals, and compensates themismatch between sub-band signals when we reconstructing the spectrum of a wideband signal by stepped-frequency chirp signals. This method estimates theamplitude-phase error and the time delay error between sub-band signals by fitting theinterferometric phase of the overlap spectra components. Besides, the window foreliminating the ghost images caused by the overlap spectra is proposed.3. The development trend of SAR imaging is to simultaneously obtain the HighResolution Wide Swath (HRWS) SAR image and moving target indication. However,spaceborne single-channel SAR system is restricted by the minimum antenna area. In order to meet the qualification of wide swath with none range ambiguity, the pulserepetition frequency can not be too high. But in order to meet the qualification of highrange resolution, the system must operate at high pulse repetition frequency. So it isdifficult to simultaneously obtain wide swath and high range resolution. In view of theabove problems, this dissertation discusses in detail how to use the multi-channel SARsystem for moving target detection in the presence of Doppler ambiguity case. A newmethod based on the space-time spectrum of GMTI for spaceborne multi-channelHRWS SAR system is presented, which is according to the space-time spectrumrelationship between moving target and clutter to contruct azimuth matching function.Firstly, the method utilizes beamforming for clutter suppression after azimuthmatching filtering, and then estimates moving target direction by fitting for the slantrange of the moving target. Secondly, focusing for the clutter suppressed data isperformed to obtain ambiguous images of the moving target, then all ambiguousmoving targets are obtained by Constant False Alarm Rate (CFAR) detectiontechnology. Finally, this method detects real targets according to the spatialrelationships of fuzzy images and motion direction estimated.4. The development trend of SAR is to simultaneously obtain high resolution inazimuth and High Resolution Wide Swath (HRWS). Based on the GMTI technology, amotion parameter estimation method for spaceborne multi-channel HRWS SAR systemsis presented. Firstly, the space time adaptive processing is adopted for cluttersuppression, then the motion trail and true azimuth position of moving target can beobtained after range compression for clutter suppression data. Secondly, focusing isprocessed. The real target is decided from ambiguous images by fuzzy Dopplerrelationship, and the azimuth shift of moving target caused by speed along the line ofthe sight direction is obtained. Finally, the motion velocity along the line of the sightdirection of moving target is obtained. The advantages of the method are its lowcomputational burden and high accuracy.更多还原