【作者】 束宇翔；

【导师】 廖桂生；

【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】 合成孔径雷达（SAR）技术与地面运动目标检测（GMTI）技术相结合，可以在获得高分辨地面场景图像的同时完成地面动目标的检测及测速定位，具有重要的民用和军用价值。相比单通道SAR系统，多通道SAR的优势在于可以利用多个方位自由度有效的抑制主瓣杂波，改善动目标的检测性能，因而受到各国学者的广泛关注。本文针对分布式卫星多相位中心SAR实现GMTI功能，开展了关于沿航迹基线形式确定、时变混合基线预处理、与高分辨率宽测绘带成像兼容设计、模糊杂波抑制以及对误差稳健的动目标参数估计等关键技术研究。主要内容包括：1、分析了多通道SAR系统中多个相位中心之间可能存在的时空关系，以交替收发模式为例，建立了多通道动目标信号模型。通过理论推导明确了形成GMTI基线的必备条件，得到如下结论：不同时间处于不同方位的两个相位中心之间的沿航迹基线，等于以同一时间基准对这两个相位中心位置进行插值后得到的等效位置之间的沿航迹差。该结论为不同通道构成方式下沿航迹基线的确定提供了依据。通过对比同一系统采用不同天线孔径划分和交替收发方式下的GMTI性能，说明了合理选择天线孔径划分和交替方式对提升GMTI性能的重要性。2、提出了分布式卫星SAR-GMTI系统时变、混合基线的预处理方法。针对分布式卫星系统中星间基线时变导致SAR图像对相干性下降的问题，从分布式卫星SAR系统的观测几何构型出发，采用总体最小二乘法拟合航迹方程，并依据方程参数重构虚拟的新平行航迹。通过补偿原航迹和新航迹上的观测波程差，获得等效的固定基线回波数据。在此基础上，针对分布式卫星SAR系统中混合基线和雷达斜视导致沿航迹基线难以精确计算的问题，推导了存在混合基线和斜视情况下沿航迹基线的计算公式，提出了基于星历和回波数据的沿航迹基线计算方法。仿真实验结果表明，时变基线补偿后，能有效提高动目标检测性能；用所提方法计算得到的沿航迹基线长度作为测速定位的基线参数，可以获得高精度的测速定位结果。3、采用方位多通道抑制多普勒模糊是实现高分辨率宽测绘带（HRWS）成像的有效方式，如果这些自由度用于主瓣杂波抑制则可以实现GMTI功能。因此，在多通道SAR系统中，有望实现两种模式的兼容工作。但此时GMTI性能会受到方位降采样引起的多普勒模糊的影响。为分析这一影响，推导了存在多普勒模糊下的杂波和动目标的信号形式，给出了系统参数的设计考虑。结合盲速、空域模糊和动目标积累损失等因素，重点讨论了在多普勒模糊情况下，脉冲重复频率（PRF）对GMTI性能的影响，并给出了一个PRF的优化选择仿真示例。根据理论分析和仿真实验结果可以得到一个选择PRF的重要原则，即应避免抑制模糊杂波的凹口因为空域模糊而折回至感兴趣速度区间。实测数据的处理结果验证了理论分析的正确性。4、提出了多特征联合挑选样本的多通道SAR模糊杂波抑制方法。针对强弱对比明显的非均匀场景中模糊强杂波的抑制问题，分析推导了模糊杂波的多通道导向矢量，以邻近强陆地杂波的近岸水面区域为例，建立了SAR图像域多通道信号模型。以模糊杂波导向的推导结果和建立的模型为依据，联合模糊杂波的幅度/相位直方图统计特性和边缘检测挑选训练样本，并采用自适应匹配滤波方法抑制杂波。仿真实验验证了所提方法的有效性。5、实际处理环境中，图像配准误差和通道相位误差不可避免，这将导致多通道SAR-GMTI处理时动目标径向速度估计和定位精度下降。针对这一问题，提出了基于自适应数据重构和子空间投影的动目标径向速度估计方法。在联合像素信号模型下，利用抑制杂波的维纳权矢量自适应重构包含动目标单元的多通道数据，再采用子空间投影算法从重构的包含动目标的单次快拍数据中估计动目标的径向速度。仿真实验和某机载多通道SAR-GMTI实测数据的处理结果验证了所提方法的有效性和稳健性。更多还原

【Abstract】 Synthetic aperture radar (SAR) combined with ground moving target indication(GMTI) can provide high resolution images of the stationary scene as well as thedetection and relocation of ground moving targets simultaneously, which has beenwidely used in many civilian and military applications. Compared with single channelSAR systems, multichannel SAR systems have the advantage that the cluttersuppression can be achieved by using the spatial degrees of freedom in azimuth, thusarousing extensive attention of the scholars in various countries. This dissertationaddresses some key techniques of GMTI for distributed satellite SAR systems withmultiple phase centers, such as the determination of the along-track baseline, the hybridand time-varying baseline preprocessing, the system design to make the GMTI becompatible with high-resolution wide-swath (HRWS) imaging, the ambiguous cluttersuppression and robust target motion parameter estimation. The main work can besummarized as follows:1. The possible temporal and spatial relationships between different phase centersin multichannel SAR systems are analyzed. The signal model of moving targets inmulti-channel SAR systems with alternate transmission and reception is established.Based on the theoretical derivation, the necessary condition for the existence of thealong-track baseline is indicated. And we can draw a conclusion that the along-trackbaseline between different phase centers whose azimuth positions are different fromeach other at different time is equal to the azimuth spacing between them after usinginterpolation to make them appear at the same time. This conclusion provides the basisto determine the along-track baseline for various multichannel formations. The GMTIperformance comparison with different antenna partitions and alternating strategiesshows that an appropriate antenna partition and alternating strategy is important inimproving the GMTI performance.2. A preprocessing method to deal with the hybrid and time-varying baseline indistributed satellite SAR-GMTI systems is proposed. The time-varying baseline makesthe correlation of SAR images decreases; and the hybrid baseline and squint make itdifficult to estimate the along-track baseline accurately. For the purpose ofcompensating the time-varying baseline, based on the geometrical configuration ofdistributed-satellite SAR systems, the total least squares algorithm is applied to fit thetrack equations of the satellites. And a new virtual parallel track is constructed with theparameters of the track equations. An equivalent constant baseline is obtained by the compensation of the path difference between the new and the original tracks. Then, theformula for calculating the along-track baseline in the case of hybrid baseline and squintis deduced, and an along-track baseline determination method based on ephemeris dataand radar echoes is proposed. Simulation results illustrate that the performance ofground moving target indication can be improved by the time-varying baselinecompensation and accurate velocity estimation and relocation of moving targets can beobtained by using the proposed method to determinate the along-track baseline.3. Using multichannel in azimuth to suppress the Doppler ambiguities allows forHRWS SAR. If the degrees of freedom in azimuth are used for the clutter suppression,the GMTI can be achieved. Therefore, a spaceborne multichannel SAR system has thepotential to offer GMTI and HRWS imaging capabilities simultaneously. However, inthis case, the GMTI performance may suffer from the Doppler ambiguity caused by theundersampling. To investigate the impact of Doppler ambiguity on GMTI performance,the multichannel signal models of the clutter and the moving target with Dopplerambiguity are derived in complex image domain. And the system design considerationsare given. Taking into account the blind velocity, the spatial ambiguity and thesignal-to-noise ratio (SNR) loss, the influence of pulse repetition frequency (PRF) onthe GMTI performance is dicussed in detail. Then a simulation example of PRFselection to optimize the system performance is provided. According to the theoreticalanalysis and the simulation example, an important principle is obtained that for the PRFselection, it is necessary to avoid the ambiguous clutter notches falling into the radialvelocity interval of interest due to the spatial ambiguity. Finally, the theoreticalinvestigations are verified by real data experiments.4. A new ambiguous clutter suppression method based on using multiplecharacteristics to select training samples with azimuth multichannel SAR systems isproposed. To deal with the ambiguous clutter suppression issue for a heterogeneousregion which contains two adjacent terrains having large reflectivity difference, such asthe near-shore water regions, the multichannel steering vector of the ambiguous clutteris derived, and the multichannel signal model in the image domain is established. Basedon the derived steering vector and the established signal model, the statisticcharacteristics of the magnitude and interferometric phase histograms combined withthe edge detection are used to select training samples, and then the adaptive matchedfiltering technique is used suppress the ambiguous clutter. The validities of the proposedmethods are verified by the simulation results.5. In practice, inevitable image coregistration error and channel phase mismatch will significantly degrade the estimation performance of the target radial velocity in theGMTI processing with multichannel SAR systems, making it difficult to relocate themoving target accurately. To overcome the influence of these non-ideal factors, a newradial velocity estimation method using the subspace projection algorithm with adaptivedata reconstruction is proposed. Based on the joint-pixel signal model, the Wienerweight vector is used to reconstruct the multichannel data vector of the pixel containingthe moving target. Then, the subspace projection algorithm is adopted to deal with theradial velocity estimation with this reconstructed single―snapshot‖data. The validityand robustness are verified by both simulations and real SAR data experiments.更多还原