【作者】 张冬晨；

【导师】 王东进；

【作者基本信息】 中国科学技术大学 ， 电磁场与微波技术， 2009， 博士

【摘要】 干涉式逆合成孔径雷达(InISAR)成像是近年来发展起来的一种将干涉技术与逆合成孔径分辨相结合的雷达三维成像方法。由于InISAR能够实现对远距离运动目标全天候、全天时的三维成像,因此在军事和民用领域都呈现出广泛的应用前景和实用价值。但是,目前在针对复杂运动目标的InISAR三维成像的实现方法上,还存在众多问题亟待深入研究和完善。本文紧密围绕InISAR三维成像的实际应用,就几类典型运动目标的InISAR三维成像实现方法展开了相关研究和探讨:论文首先研究了InISAR三维成像的图像配准问题。干涉天线间的图像配准是InISAR三维图像恢复的关键步骤。在深入分析干涉图像失配原因的基础上,提出了两种图像配准的方法:其一,在目标存在独立特显点的情况下,对各天线进行基于单特显点的初相校正,在完成运动补偿的同时实现了以该特显点为聚焦中心的配准处理;其二,当目标不存在特显点时,在完成各天线的平动补偿后,通过估计目标散射中心沿两基线方向的角运动参数,构造补偿相位来实现图像配准。仿真结果表明,这两种方法都能够获得良好的配准效果,为进一步实现InISAR三维成像奠定了基础。其次,论文对高速运动目标的InISAR三维成像方法进行了研究。目标的高速运动会导致距离色散现象,造成ISAR成像散焦,进而严重影响干涉处理的精度。因此,论文提出了两种有效解决距离色散影响的高分辨InISAR三维成像实现方法:第一种方法从速度补偿的思路出发,提出了基于RWT的分步速度估计方法,通过精确估计目标的径向速度并对回波进行补偿来实现三维成像;第二种方法基于瞬时距离像的思路,通过构造高速运动目标回波的时频域图像切片(即瞬时距离像),消除了距离色散的影响,进而实现InISAR三维成像。仿真结果表明,上述成像方法都能够正确的实现对高速运动目标的InISAR三维成像。然后,论文研究了机动目标的InISAR三维成像方法。与高速运动目标回波的脉内距离色散不同,目标在成像期间的机动特性会使回波序列的横向多普勒呈现多分量LFM信号的特点,造成传统RD成像的横向聚焦失效。本文首先在距离-瞬时多普勒(RID)成像的思路下,提出一种基于FRFT线性时频分析的机动目标三维成像方法,在分数阶域完成机动目标散射点的横向分辨,进而实现三维成像。针对双线性类时频方法由于相位信息丢失特性而不能应用于机动目标InISAR成像的问题,本文进一步提出了一种基于多天线联合互时频分布的InISAR三维成像方法,不仅克服了双线性时频变换不能在InISAR中应用的限制,而且能够在一次天线间的互时频变换中同时实现相同距离单元内各散射点的横向分辨和相位干涉处理,实现了机动目标的InISAR三维成像。仿真结果表明,上述两种InISAR成像方法都能够有效的获得机动目标的高分辨三维图像。最后,论文对大尺寸运动目标的InISAR三维成像方法进行了研究。大尺寸目标的高分辨成像不满足传统ISAR的微动模型近似,其边缘散射点的MTRC现象对运动补偿和横向分辨两个环节造成严重的影响,无法形成清晰的三维成像结果。论文首先提出一种结合积累相关与曲线拟合的运动补偿方法,在进行成像校正前,从整体上提高了在包络相关性下降情况下的运动补偿的精度;然后提出一种在InISAR多天线成像结构下的反投影转角估计方法,能够利用三维成像的优势获得转角的精确估计,最后结合极坐标格式算法,实现了大尺寸目标的高分辨InISAR三维成像。更多还原

【Abstract】 Interferometric inverse synthetic aperture radar(InISAR) imaging is a recently developed three-dimensional(3-D) imaging method which combines interferometric technology with the inverse synthetic aperture radar.InISAR is able to carry out 3-D imaging of a far-field moving target under all-weather condition and any required time,so it has shown a wide range of prospects and practical value in the field of military and civilian applications.But at present,the realization of InISAR 3-D imaging method of complex moving targets are still facing many problems and need in-depth researching and improving.This dissertation has been launched to study and investigate the InISAR imaging methods of three typical moving targets,which are valuable for the practical application.First of all,the image registration problem of InISAR is studied.Image registration between interferometric antennas is a crucial step for InISAR 3-D image restoration.Based on the analysis of the image mismatch,two effective image registration methods which make use of the characteristics of different targets are proposed.The first method is that if the target contains only one isolated scatterer,we carry out single isolated scatterer phase-adjustment of each antenna independently to achieve image registration,which will synchronously consider the single isolated scatterer as focus center.The second one is that when there are no isolated scatters exist,after uniform radial-motion compensation,we construct compensatory phase and achieve image registration by estimating angle motion parameters of target scattering centers along two-dimensional baseline direction.Simulation results show that both methods can obtain accurate registration results.Secondly,the 3-D InISAR imaging method of high speed moving target is studied.The high radial velocity between radar and target will result in a range dispersion phenomenon which will defocuses ISAR image and seriously affect the accuracy of the interference restoration.Therefore,two effective high-resolution InISAR 3-D imaging methods for high speed moving target are proposed.One is based on the approach of velocity compensation,a precise velocity estimation method based on RWT is proposed and 3-D image can be achieved through the compensated echo.Unlike the upside approach of speed compensation,the second method is based on instantaneous range profile,which can eliminate the dispersion effects in the time-frequency slice and directly used for 3-D InISAR imaging.Simulation results show that the above-mentioned imaging methods can both achieve 3-D InISAR imaging of high-speed moving targets correctly.Thirdly,the 3-D InISAR imaging method of maneuvering targets is studied. Unlike the range dispersion caused by high speed,Target’s maneuvering characteristic during imaging time will make the cross-range Doppler becoming multi-component LFM signals,by using traditional RD imaging method will cause the defocus of cross-range imaging.Under the idea of range- instantaneous Doppler(RID) imaging, a 3-D InISAR imaging method of maneuvering targets based on FRFT linear time-frequency transform is proposed firstly.Scatterers’ cross-range separation can be achieved in the fractional domain and then a high-resolution 3-D image can be reconstructed.Aiming at the problem of the restrictions of bilinear time-frequency analysis method which can not retain the interferometric phase,we further proposed a 3-D InISAR imaging method based on joint cross-time-frequency distribution.This multi-antenna-based method not only can overcome the restrictions of the bilinear time-frequency transform,but also can achieve scatterers’ cross-range reparation and phase interferometfic processing in one transform at the same time,which can directly used to reconstruct 3-D InISAR image.Simulation results show that both two proposed methods can effectively obtain the high-resolution 3-D InISAR image of maneuvering target.Finally,the 3-D InISAR imaging method of large size target is studied. High-resolution imaging of large size targets are not satisfied with the traditional ISAR micro-rotating approximation,the MTRC phenomenon of scatterers at the edge both seriously effect motion compensation and cross-range resolution which will decline the accuracy of InISAR imaging.A method of motion compensation which combines accumulation correlation with the curve fitting is proposed at first to increase the accuracy of motion compensation under MTRC condition.Then,the back-projection rotation angle estimation technique in the structure of multi-antenna InISAR imaging is proposed.This technique can obtain accurate estimation of the rotation angle taking the advantages of 3-D imaging,then combing with polar format algorithm,a clear 3-D InISAR image of large size target can be formed at last.更多还原