【作者】 杨桃丽；

【导师】 保铮；

【作者基本信息】 西安电子科技大学 ， 信号与信息处理， 2014， 博士

【摘要】 星载合成孔径雷达（Synthetic Aperture Radar, SAR）因其全天时、全天候的全球观测能力，受到了越来越多国家和地区的重视，在军事侦察、国民经济建设和科学研究中得到了广泛的应用。但是，传统星载单通道SAR存在最小天线面积限制，无法同时满足高分辨宽测绘带（High Resolution Wide Swath, HRWS）成像的要求：方位高分辨要求较高的脉冲重复频率（Pulse Repetition Frequency, PRF），而宽距离测绘带则要求较低的PRF。多通道结合数字波束形成（DigitalBeam-Forming, DBF）技术可克服这一限制，从而实现高分辨宽测绘带SAR成像。本文针对星载多通道HRWS SAR系统，重点研究了成像处理的几个关键技术。全文总体上分为两个部分：第一部分主要研究了星载方位多通道高分辨宽测绘带SAR成像处理技术，方位多通道SAR是目前高分辨宽测绘带成像实现最多的体制，在保证距离测绘带宽和距离模糊度要求的前提下，通过多通道DBF处理实现多普勒模糊抑制从而得到HRWS SAR图像。随着分辨率和测绘带宽的提高，回波数据量也大大增加了，对星上存储设备和传输链路等也提出了更高的要求，第二部分针对此问题，研究了星载距离多波束HRWS SAR成像技术，该技术在保证其它性能参数的前提下，可大大降低回波采样数据量，为未来星载HRWS SAR的实现提供了新的方案。本文的主要工作可概括如下：1.星载方位多通道HRWS SAR成像处理技术本文第二章和第三章首先分析了星载方位多通道HRWS SAR回波信号模型，给出了两种典型的成像处理方法，并针对星载方位多通道HRWS SAR成像的特点，开展了以下研究工作：针对星载方位多通道HRWS SAR系统，推导了三维坐标系下接收通道的等效相位中心相位补偿公式。现有的方位多通道HRWS SAR成像算法均假设各通道接收回波补偿一常数相位后可等效为参考接收通道的延时，但并未给出具体的补偿公式，或只给出了两维坐标系下的补偿值，也未考虑发射通道和接收通道间的垂直航向基线。本文所提方法充分考虑了发射通道与接收通道间的三维空间位置关系，同时补偿了由沿航向基线和垂直航向基线引起的等效相位中心相位值，并对残余相位误差进行了分析，指出当接收通道与参考接收通道间存在垂直航向基线时，相位补偿值存在一定的空变性。当空变引起的相位误差不可忽略时，可在距离压缩后利用先验数字高程模型（Digital Elevation Model, DEM）辅助分块补偿。计算机仿真实验验证了本文所提方法的精确性。针对方位多通道HRWS SAR系统，对空时自适应处理（Space TimeAdaptive Processing, STAP）法的处理性能进行了分析。传递函数法和空时自适应处理法是目前两种典型的多普勒模糊抑制算法，前者已有大量文献对其处理性能进行了分析，并利用地基、机载和星载实测数据对其进行了验证，却鲜有文献对STAP法的处理性能进行分析。基于此，本文首先从理论上分析了利用STAP进行多普勒模糊抑制后的成像等效相位中心位置，验证了STAP的保相性和保幅性，经多普勒模糊抑制后输出回波可看作参考接收通道增加脉冲重复频率后得到的无模糊回波，且各个方位时刻回波所对应的卫星轨道位置由参考接收通道的位置决定，这为后续的干涉处理和目标定位奠定了基础。除此之外，本文还从不同于现有文献且更利于理解的角度分析了STAP解多普勒模糊后的信噪比损失和方位模糊信号比，并利用仿真实验对其进行了验证。实验表明，当PRF偏离均匀采样时，相比其它模糊抑制算法，STAP的处理性能更优，能更好地保留回波信号能量，抑制多普勒模糊。针对方位多通道SAR系统，分析了通道误差因素及其影响。利用DBF技术进行多普勒模糊抑制要求各通道间的特性一致，但在实际情况中，由于加工工艺、运行环境等的影响，通道间不可避免地存在误差，此外，受测量仪器精度的限制，通道位置也存在测量误差。基于此，我们首先对通道误差因素进行了分析，根据各误差因素对DBF的影响将其归结为通道幅度误差、通道相位误差和通道沿航向位置误差，然后推导分析了通道误差对HRWS SAR成像的影响，并利用计算机仿真实验对其进行了验证。实验表明，通道相位误差对DBF的影响最大，而通道沿航向位置误差的影响相对较小，但也应控制在厘米量级，通道幅度误差可通过简单的通道均衡予以消除。与其它算法相比，STAP法具有更高的误差容忍度。提出了两种方位多通道HRWS SAR系统通道误差估计和补偿方法。在实际情况中，通道误差不可避免，为了提高HRWS SAR成像性能，必须对其进行补偿，特别是通道相位误差。针对方位多通道SAR系统，我们提出了两种通道相位误差估计方法：信号子空间法（Signal Subspace Comparison Method, SSCM）和天线方向图法（Antenna Pattern Method, APM）。信号子空间法基于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这一特性对通道误差进行估计。首先对利用回波信号估计得到的协方差矩阵进行特征分解得到信号子空间，然后与利用系统参数得到理想的信号子空间相比较，从而得到通道相位误差，与其它算法相比，该方法运算量更小，且适用范围广。天线方向图法假设观测场景均匀分布，在此条件下，利用发射和接收天线方向图对理想信号导向矢量进行加权，然后与回波信号协方差矩阵对比，得到各通道间的相对相位误差。天线方向图法无须特征分解，运算量小，但适用范围受限。最后利用地基实测数据对两种方法的有效性进行了验证。2.星载距离多通道HRWS SAR成像处理技术本文第四章给出了一种新的高分辨宽测绘带SAR成像技术。随着分辨率和测绘带宽的提高，回波信号的数据量大大增加，对星上存储设备和传输链路的要求也随之提高。基于此，我们针对距离多波束分时发射技术，给出了详细的系统实施方案和处理方法，并对其距离模糊度（Range Ambiguity to Signal Ratio, RASR）等系统性能进行了分析。通过由远及近分时发射信号，各子波束回波将重叠在一起，这样可大大缩短接收窗长度，减小回波采样数据量。然后利用DBF技术进行子波束分离，最后采用传统SAR成像方法即可得到高分辨宽测绘带SAR图像。当观测场景存在地形起伏时，可借助先验DEM进行子波束分离。实验表明，现有的DEM精度（例如SRTM DEM的精度约为17m）所引起的误差可忽略。最后利用计算机仿真实验验证了本文方法的有效性。更多还原

【Abstract】 Spaceborne synthetic aperture radar (SAR) has been receiving more and moreattention because of its cloud-penetrating and day and night operational capabilities.Nowadays, SAR is widely used in fields of military reconnaissance, civil constructionand science research. However, traditional spaceborne single channel SAR systemsuffers from a tradeoff between the achievable resolution and swath width, i.e. theminimum antenna constraint. Fine azimuth resolution requires high pulse repetitionfrequency (PRF), while low PRF is utilized for wide swath. Fortunately, incorporatedwith digital beam-forming (DBF) processing, multi-channel spaceborne SAR systemsare able to overcome this limitation and yield high resolution and wide swath (HRWS)images.In this dissertation, some of key techniques for multi-channel spaceborne HRWSSAR system have been studied. The whole dissertation is composed of two main parts.In the first part, multi-channel spaceborne SAR system in azimuth is studied, which isone of the most typical systems for high resolution and wide swath SAR imaging. Insuch SAR systems, the low PRF, usually much lower than the instantaneous Dopplerbandwidth, is employed to avoid the range ambiguities and implement wide swath,which results in the ambiguous Doppler spectrum, and DBF is utilized to suppressDoppler ambiguity yielding HRWS SAR images. With the improvement of bothresolution and swath, the amount of raw data is increased greatly, imposing higherrequirement for satellite storage and transimission link. For such problems, multipleelevation beam technique for HRWS spaceborne SAR systems is studied. Thistechnique can substantially reduce the amount of data to be recorded and stored on thesatellite without deteriorating other performances, and provide a probable scheme forfuture spaceborne HRWS SAR imaging.The main work of the dissertation is summarized as follows:1. Azimuth multi-channel spaceborne HRWS SAR imaging techniquesIn Chapter2and Chapter3, the signal model of multi-channel spaceborne HRWSSAR in azimuth is analyzed, followed by two typical imaging methods. For the specialcharacteristics of azimuth multi-channel spaceborne HRWS SAR system, the followingworks have been done.A common echo model based on three dimensional coordinates for spacebornemulti-channel SAR systems is built with the consideration of both along-track and across-track baselines between transmitter and receivers, as well as the effective phasecenter (EPC) phase compensation equation. Almost all the methods for the azimuthmulti-channel SAR imaging assume that the echoes received by each channel can beregarded as that received by the reference channel with an along-trackbaseline-dependent time delay after certain phase compensation. The phasecompensation, however, has not been given in a general case. Besides, neither theacross-track baseline between transmitter and receivers is considered, which is of greatsignificance to distributed high resolution InSAR systems, nor the sensor orbitinformation after focusing is given, which is the basis for the interferometry and targetlocation. In this dissertation, a common EPC phase compensation method based onthree dimensional coordinates of transimitter and receivers is given, considering thephase terms arising from both the along-track and across-track baselines are given, andthe residual phase error is analyzed too. The experiements show that when cross-trackbaseline between the reference channel and other receivers exists, the compensatedphase is varied with range swath and target elevation. If the phase error caused by spacevariance cannot be ignored, the data can be compensated with the assistance of coarsedigital elevation model (DEM) after range compression. The computer simulationconfirms the accuracy of the method.Regarding the multi-channel spaceborne SAR system in azimuth, theperformance of the space time adaptive processing (STAP) approach applied to HRWSSAR imaging is investigated. The multi-channel transfer function method and theSTAP-based approaches are two typical algorithms for suppressing Doppler ambiguities.The performance of the reconstruction algorithm has been well analyzed in variousliteratures, and demonstrated by the ground-based, airborne and spaceborne campaigns.While, there is no literature discussing the performance of STAP-based approach indetail. In this dissertation, the phase preservation of STAP is confirmed by analyzingthe position of imaging EPC of output data after Doppler suppression. It shows thatafter Doppler ambiguity suppression, the echo can be regarded as unambiguity ones thatobtained by the reference channel with higher PRF, and the radar positioncorresponding to each azimuth sample time is determined by the reference receiver. Thephase preservation of SAR imaging guarantees the following interferometry processingand target locating etc. Besides, two important parameters, signal to noise ratio (SNR)scaling and azimuth ambiguity to signal ratio (AASR), are evaluated, and comparedwith the multi-channel reconstruction method. The derivations of SNR scaling andAASR here are more legible and comprehensible than those introduced in other literatures. The numerical analysis is confirmed by the simulated results, which showsthat the STAP-based method has a better performance than other methods when PRFdeviates from the uniform sampling.The channel errors and their effects are analyzed for multi-channel SAR systems inazimuth. Adopting DBF technique to suppress Doppler ambiguity requires that thecharacteristics of each channel are identical. In practice, however, for someenvironmental factors, the channel errors are unavoidable. The mismatch amongchannels will degrade the performance of DBF. In addition, the limited precision of themeasurement will also cause errors. The channel error factors are analyzed in thisdissertation and then decomposed into channel gain, phase and along-track positionerrors according to their effects on DBF. The influence of channel errors on HRWSimaging is analyzed in detail and confirmed by computer simulation experiments. Theresults show that the effect of channel phase error is great, the effect of the channel gainerrors can be compensated by simple channel balancing, and the influence of thealong-track position errors is little. The STAP-based method has a better ambiguitysuppression performance than the reconstruction approach because of its capability ofplacing nulls in the directions of the interferences.Two novel methods are proposed to estimate channel errors for multi-channelHRWS SAR systems in azimuth. In practice, the channel errors are inevitable andshould be compensated in order to improve the performace of HRWS SAR imaging.Therefore, two channel error estimation methods for azimuth multi-channel SARsystem are presented: the Signal Subspace Comparison Method (SSCM) and AntennaPattern Method (APM). SSCM is based on the fact that the space spanned by the signaleigenvectors is equal to that by the practical steering vectors. The signal subspace isobtained by eigen-decomposing the echo covariance matrix, and then compared with thetheoretical signal subspace that derived from the system parameters to obtain the phaseerror. This method has great advantages of light computational load and high accuracy.Furthermore, it has no requirement that the SAR systems must operate in rightside-looking mode. The APM incorporates with the antenna patterns to estimate thechannel errors directly without matrix decomposition and inversion processing, which isvery efficient, but only suitable for uniform distributed scenes. Both the theoreticalanalysis and experiments demonstrate the effectiveness and efficiency of these twomethods.2. Elevation multi-channel spaceborne HRWS SAR imaging techniques In Chapter4, the novel HRWS SAR imaging technique is proposed. With theimprovement of resolution and range swath, the amount of data is increased greatly andsubsequently impose strict requirement on satellite storage and transmission link. Inorder to solve such problem, the detailed system design scheme and processing methodbased on multiple elevation beam technique for HRWS imaging is presented as well asits system performance such as range ambiguity to signal ratio (RASR). The wideimage swath is illuminated by a sequence of narrow and high-gain antenna beams. Thewhole antenna plane transmits a narrow beam to illuminate a far-range and subsequentlyproceed to the near range, and all the channels receive the radar echoes simultaneously.As a result, the echoes from different subswaths will overlap in the receivers, therebythe amount of data to be recorded and stored on the satellite will be reduced. All theechoes from different subswaths can be separated from each other by DBF and yieldingthe wide swath SAR image. For terrain with strong topographic variance, the subbeamscan be separated with the assistance of coarse DEM. The experiment shows that theerror caused by DEM accuracy is neglectable. Finally, the simulation data confirms thevalidity of the method.更多还原