【作者】 王亮；

【导师】 何建国； 周智敏；

【作者基本信息】 国防科学技术大学 ， 电子科学与技术， 2007， 博士

【摘要】 机载低频超宽带合成孔径雷达(UWB SAR)具有穿透叶簇和浅地表实现对隐蔽目标成像的能力,能够对宽幅面、大场景区域进行远距离侦察,军事应用潜力巨大。本论文以国防预研项目为背景,针对现有实测机载P波段UWB SAR数据成像处理中所面临的问题,围绕宽幅成像算法、运动补偿和图像干涉三个方面展开了深入研究,取得了有益的成果。在宽幅UWB SAR成像算法方面,论文着重研究了SAR成像处理中的相位近似问题,提出了两种P波段宽幅UWB SAR成像的一致聚焦子测绘带划分方法:残余耦合相位法和距离谱特性法。为将理论研究结果实用化,论文还研究了UWBSAR中非正交旁瓣的产生原因和加变口径窗抑制的方法,加窗后的回波有更大的耦合相位容忍度,增加了一致聚焦测绘带的宽度,所得结论可用于指导宽幅UWBSAR分段成像。在UWB SAR运动补偿方面,论文分为低阶运动补偿和残余运动补偿两章来研究。为适应UWB SAR子孔径模型的多普勒参数距离空变性,提出了距离向分块估计多普勒调频率,联立稀疏矩阵方程求解Y方向加速度的算法。针对实测UWBSAR回波多普勒参数估计中的几个问题,给出合理分析与解释,提出图像偏置算法实现的几点经验手段和确定子孔径大小的方法。为满足UWB SAR子孔径残余相位误差模型的特性,提出以估计两个空不变因子为目的的加权相位梯度自聚焦算法。上述研究成果已由实测机载UWB SAR数据验证。在UWB SAR图像干涉方面,论文首先从原理上分析了UWB InSAR与一般窄带InSAR相比的优势所在,总结了UWB InSAR技术的难点。然后以实现一段实测重航过机载UWB SAR图像的干涉为目的,研究了图像配准、预滤波和干涉相位图估计等有关问题。其中,提出了基于局部自适应自相关函数估计的二维实图像插值算法解决二维复图像插值问题;提出对联合自相关矩阵进行加权处理的措施,改进了基于联合子空间投影的干涉相位图估计算法;分析了时变基线形成方位向平地效应的原理,推导出干涉相位方位谱峰值偏移量公式。最后完成了UWBSAR图像数据的干涉,得到与实际地形情况的理论预测结果相吻合的干涉相位图。更多还原

【Abstract】 Airborne ultra-wideband synthetic aperture radar (UWB SAR) working in low frequency band can produce fine images of concealed targets with the capability of penetrating the foliage and the earth surface to reconnoitre wide-range scene, which is very valuable for military purpose. This dissertation work was supported in part by the National Defense Research Projects. According to the problems currently faced in the real data image formation processing on airborne P-band ultra wideband synthetic aperture radar (UWB SAR), this dissertation researched three subjects—wide-range scene imaging algorithm, motion compensation and image interferometry.On the first subject, the phase approximation of image formation algorithms for UWB SAR operating on P band was discussed in detail. The methods to determine the consistent focusing swath width in Chirp Scaling algorithm using residual coupling phases and range spectrum characteristics were also proposed. In order to make the theoretic results more realizable, the dissertation analyzed the formatting non-orthogonal sidelobes in UWB SAR images. The weighting function with variable length was employed to give a suppression of non-orthogonal sidelobes. By this method, echoes can be more tolerant to coupling phases in focusing so that the consistent focusing swath width was greatly increased. Using the above conclusions as guidance in segmented imaging, wide swath UWB SAR data had been successfully focused.In UWB SAR motion compensation (MOCO), the content was divided into two chapters, which concentrate on low-order MOCO and residual MOCO, respectively. To solve the range dependency of doppler parameters in UWB SAR subaperture model, an airborne motion error estimation algorithm was presented, in which doppler frequency modulation ratio was estimated by sub-range block data. Subsequently, these estimated values were jointed as a sparse matrix equation and the acceleration of direction Y was attained. Meanwhile, in real UWB SAR data processing, some questions from the doppler parameter estimation procedures was proposed and analyzed. And on the basis of the above work, the dissertation proposed several practical and important methods in map-drift algorithm realization and subaperture length calculation. To meet the feature of UWB SAR subaperture residual phase error model, an improved PWE-PGA (phase weight estimation - phase gradient autofocus) algorithm was proposed, whose intention was to get two range-independency factors. The effectiveness of the proposed methods had been verified with real UWB SAR data.In UWB SAR images interferometry, the dissertation first analyzed the advantages in UWB SAR interferometry (UWB InSAR) comparing to general narrow-band InSAR and concluded the difficulties in UWB InSAR realization. Afterward, the processing procedure of interferogram generation from two real UWB SAR images was demonstrated, which includes 2D-complex-images coregistration, pre-filtering and interferogram estimation. An interpolation algorithm based on local adaptive autocorrelation function estimation was proposed to appease complex-image interpolation. Based on the joint subspace projection estimating interferogram algorithm, a weighted joint autocorrelation matrix method was also proposed to improve the precision of interferogram estimation. In addition, the dissertation analyzed the time-varying baseline and deducted the offset formula of azimuthal spectrum of interferometric phase. Finally, the rationality of the interferometric phase fringes generated from those two real UWB SAR images was proved by the predicited results based on practical terrain.更多还原