【作者】 姜正林；

【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， 2001， 博士

【摘要】 随着雷达技术的发展，现代雷达的最显著特点之一就是对目标的分辨能力的提高，以宽带微波技术和先进的信号处理技术为基础的成像雷达，它作为一种高分辨率的雷达，能够对观测对象进行全天候、全天时、远距离的成像，有着广泛的应用前景，成为研究的热点。 对于ISAR成像，主要是针对空中运动目标的成像，传统的简单的RDA算法是基于散射点的微动模型，但是实际的目标并非完全符合这个要求，所以必须采用一定的措施校正因为模型失配引起成像误差，本文首先对此进行了讨论。对于现役的大量的低分辨雷达，由于纵向的低分辨率，所以只能对得到的回波数据进行横向的多普勒成像，为此本文对低分辨雷达的成像进行了系列的分析，提出了一些算法，并用实测数据和计算机仿真数据进行了验证。 同样建立在成像技术上的干涉技术日益成为雷达成像中的焦点，它不仅为目标识别和检测提供了辅助信息，并能实现简单成像不能完成的诸如三维成像、定标、动目标检测等功能，而基线则为干涉系统的关键。本文就干涉技术的应用进行了一些探讨。 本文的具体内容可归纳如下几点： ◆分析了用尺度变换可以消除由于径向速度引起的越距离单元走动，对ISAR目标进行越距离单元走动校正算法进行了讨论，并把尺度变换校正越距离单元走动后成像的方法推广到机动目标成像，也给出适合于平稳和机动大目标的成像算法流程，并通过仿真数据和实测数据的检验。这个算法流程不但适用于ISAR空中运动目标成像，如果成像之前进行了杂波抑制也适用于地面运动目标和海面舰船目标成像。 ◆分析了低分辩雷达的成像算法，建立在最小熵补偿基础上的算法得到的横向像的分辨率仍然比较低，尤其对大的目标和机动目标，为此我们采取了一阶近似的RELAX算法搜索目标的初始频率、幅度、解调率从而获得目标的动态一维像，仿真数据和试验结果证明该方法是有效的。 ◆针对现役的低分辩雷达，它的纵向分辨率通常达一百米甚至数百米对于编队飞行的目标，由于目标的间距和方位张角都小于低分辨雷达的可分辨单元，．11．低分辨ISAN成像及干涉技术应用研究一 因此直接无法分辨编队目标的架数，我们借鉴ISAR的技术，通过较长时间 的相干积累，在多普勒频域上对目标进行分辨。而对于编队目标，可分为近 似刚性的多目标和非刚性的多目标，所以对于可以近似为刚体的编队目标相 对位置固定，运动方式一致，可以近似看作一个大目标，采用最小墒准则对 平动相位的进行补偿，但是大多数并不满足刚体近似的编队目标，由于目标 在相干积累时间的多普勒频率近似呈线性变化，通过对较短数据利用RELAX 的时频分析方法，提高了频率域上目标分辨的性能。由于横向分辨率取块于 横向积累时间，所以利用Radnwigner变换得到瞬时的一维横向距离像大大 提高了对编队目标的分辨，对仿真和实测数据的大量分析结果表明此方法的 有效性和可行。为了更准确进行架次的识别，还可以进行长时间对编队目标 的动态横向一维像进行观测。 夸由于横向一维像的尺度受相干积累时间和转动角度所诀定，所以要进行有效 的目标识别，必须确定横向一维距离像的横向尺度，本章根据ISAR成像原 理，采用水平放置的两个天线进行干涉的技术，给出了一种低分辨目标一维 横向距离像定标的方法，根据理论分析和计算机仿真结果表明，该方法能够 准确的进行横向距离定标。这对进一步进行目标识别提供了有效的依据。 令对于ISAR成像的特点，由于横向是由横向相干积累得到的多普勒频率分布， 所以根据目标的运动方式不同，得到的目标的横向像有正负之分，所以给目 标识别带来了麻烦，尤其是对低分辨雷达只能得到的横向一维距离像，缺乏 二维ISAR像的直观性。分析了各种运动方式所得到的多普勒分布规律，选 取未发生角闪烁的散射点，利用归一化相位和差单脉冲测角方法来对散射点 的真实分布方向进行判断。最后通过计算机仿真结果也证明该方法的有效 性。 旁通过对干涉定位系统检测地面慢速动目标及方位确定的原理阐述，分析了干 涉系统基线对该定位系统的影响，并给出了基线的上限。更多还原

【Abstract】 With the development of radar technology, one of the obvious features of radar is the improved resolution. The imaging radar is a type of high-resolution radar, which is based on the technology of broadband microwave and advanced signal processing. By use of the imaging radar, which has broad applications and has become the focus of the field, the target can be imaged hi all weathers and at anytime from far distance.The object of IASR imaging is the moving target in space. The traditional RDA algorithm is based on the model of micro-motion scatters, but it is not fit for the real target at all. We must adopt some methods to correct the error, which is caused by the model mistermination, and the discussion is mentioned in the paper firstly.For some low-resolution radar on active duty, we can only get the Doppler cross image of the echo because of their low longitudinal resolution. Therefore some analysis and algorithms for low-resolution radar are given and verified by measured data and simulation data results.The interference technology is one of the focuses of radar imaging increasingly. Not only does it provide some additional information of target recognition and detection, but also has it other applications to three-dimension imaging, scaling and moving target detection and so on. And the baseline is one of the keys of the interferometer system. Some applications of interferometer technology are discussed.The summarization of the main work is as follows:+A method that can correct the migration through range cells caused by radial speed is presented by using scale transformation, and the method is extended to maneuvering target imaging. The algorithm flow, which is applicable to large-scale target whether it is uniform or maneuvering, is given. The results obtained by using simulation data and real data show that this method is effective. The method can hold not only for ISAR of flying target, but also for ground moving object and sea ship after adaptive clutter suppression.+The analysis of the algorithm of low-resolution radar is made. After the compensation based on the rule of the minimum entropy, the resolution of thecross-range image is lower as the large size and non-uniform rotation, which cause the image to be not right at all sometimes. In order to eliminate effect of the non-stationary rotation, the first order approximation and RELAX algorithm are adopted to search the initial frequency, amplitude, frequency modulation rate and obtain the dynamic image of the target at any time. Simulation results obtained using data and real data show that this method is effective.^For the active low-resolution radar, its longitudinal resolution usually is more than one hundred meters and as high as several hundred meters. Due to the short distance among the planes which fly in a group, the conventional low resolution radar can not distinguish them in both distance and azimuth ?If we use the technology of ISAR to resolve the difference among Doppler frequency of the targets and obtain a fine resolution cross-cross image, we may separate them, but a long time of coherent processing is needed <, For the formation targets, it can be approximately divided to rigid body and nonrigid body, so for the formation targets, that can be regarded as rigid which has a relative position and an identical movement, can be approximately considered as a large target, and be compensated by translational phase with the rule of minimum entropy , but for the most those cannot accord with the approximation of rigid, being the Doppler-frequency of the aim is linear changed, by the RELAX method with short data, increases the resolving performance of multiple target to the aim in the frequency domain, since cross-range resolution is based on the accumulative time, so it is greatly improve the resolution to formation targets by the instant cross-range image which produced by Radon-Wigner transformation.+As the scale of cross range image is dependent on the coherent accumulation and the rotation angle, we must s更多还原