【作者】 张兴敢；

【导师】 朱兆达；

【作者基本信息】 南京航空航天大学 ， 通信与信息系统， 2002， 博士

【摘要】 逆合成孔径雷达(ISAR)是一种高分辨成像雷达，能够获取更多的目标信息。国际上对ISAR的研究已经取得很多成果，但真正投入应用的报道很少。雷达目标识别是国防上急需解决的一个重要问题，解决途径之一是用ISAR对目标进行二维成像，基于目标像进行目标识别。ISAR是大带宽雷达，制造成本高，在短期内大规模应用比较困难。将ISAR技术用于窄带相参雷达，实现高的横向距离分辨(横向一维成像)，可以解决目前国防上尚待解决的目标架次分辨的问题。本文研究一维ISAR成像处理，二维ISAR实时成像以及基于ISAR二维像的目标识别，有利于推动ISAR的应用，不仅在国防上有非常重要的意义，在航空管制、港口监视等民用领域也有广阔的应用前景。 文中一维ISAR成像处理、二维ISAR实时成像、基于ISAR二维像的目标识别研究均用外场数抓进行了验证，给出了结果。论文共有七章和一个附录，内容如下： 第一章介绍ISAR二维成像和实时成像，ISAR一维成像以及目标自动识别的基本概念，并概述全文内容。 第二章研究一维ISAR成像方法，包括距离对准、相位补偿的几种计算方法、单目标成像和多目标成像。给出用相位差分计算相位补偿参数的方法。分析了近似刚性多目标成像和非刚性多目标成像的差别，给出了成像方法和外场数据成像结果。 第三章研究ISAR目标运动参数估计。提出横向一维像快速定标方法和切向速度计算方法。研究用目标移动速度和多普勒频移相结合克服测速误差大和速度模糊的矛盾。通过定标，能够准确估计目标尺寸，有利于分辨单目标和多目标或目标识别。 第四章研究目标架次识别。提出横向动态一维像的概念及其计算、显示方法，介绍基于横向动态一维像的目标架次识别方法及其相关问题。 第五章研究ISAR实时二维成像、图像后处理。讨论ISAR二维像实时成像的概念及其与SAR实时成像的差别。研究用软件实现非重叠实时成像方法。图像后处理包括图像连通处理、干扰条纹抑制、图像质心归一化等。其结果有利于图像特征提取，也符合光学图像的观察习惯。 第六章研究基于ISAR目标二维像的自动目标识别问题，研究了四种特征及其提取方法，人工神经网络的识别过程。外场数据验证表明，所述目标识别方法效果良好。 第七章对全文进行了总结，提出有待深入研究的工作 附录介绍384雷达用作一维ISAR的外场数据采集方案。研究数据采集系统正交误差、直流偏移、系统干扰和噪声的计算方法。给出外场数据分析结果。更多还原

【Abstract】 Inverse synthetic aperture radar (ISAR) is a kind of high resolution imaging radar. It can take more message than narrow-band radar. Though many outcomes have been taken out on imaging research, the reports number of applying the ISAR imaging in practice are very few. The recognition tecnique of radar targets is a urgent problem in national defence. One of the routes of recognition targets is imaging to target with ISAR. ISAR is a wide-band radar, so its price is very expensive. It is very difficult to widely use the ISAR in the near future. The tecnique of ISAR applies to nanrrow-band radar so that it has high accros-range resolution (ID imaging) to tell targets apart in nantional defenc. This thesis studies on ID imaging processing, ISAR 2D imaging in real time, and object imaging recognition. Those would extend ISAR and have important meaning not only in national defenc but also in civil.ISAR ID imaging processing, ISAR 2D imaging in real time, and object imaging recognition have been verified with field data. This thesis consists of six chapters and an appendix. The content is as follows:Chapter 1 introduces the essentials of ISAR ID imaging, ISAR 2D imaging in real time and object imaging recognition based on ISAR imaging.Chapter 2 deals with ISAR ID imaging, contains the algorithms of the range realignment, the phase compensation, single target imaging and multi-targets imaging. The method of difference phase compensation in ID ISAR is given out. The difference between quasi-inflexibility targets imaging and non-inflexibility targets imaging is analyzed. The imaging results of field data are given out.Chapter 3 studies estimation of moving target parameters. The methods of Fast ISAR ID imaging scaling and calculating the across speed of target are proposed. The distance moving rate and Doppler frequency shifting are gathered to correct the dim speed to decrease the speed error. After scaling, the scale of a target can be estimated correctly for targets resolution or target recognition.Chapter 4 proposes the idea of ISAR dynamic across 1D imaging. The calculation method and display mode of the imaging are introduced. The method of resolving targets based on the imaging and its relative problems are introduced too.Chapter 5 studies ISAR imaging in real time and image post-processing, and discusses the idea of ISAR imaging in real time and the difference with SAR imaging in real time. It introduces the method of non-overlapping ISAR imaging in real time based on software. The image post-processing includes image connectivity processing, interference stripe suppression and centroid normalization. The results are advantageous to theextraction features, and look as if optical imaging.Chapter 6 studies ATR based on ISAR imaging. The content includes explanation of four features of a target image, the algorithms of extraction features, and recognition target based on artificial neural network algorithm. The experimental results with field data verify that the methods of recognition target are efficient.Chapter 7 summrizes the whole work of this thesis and show the researches to be carried out further.Appendix introduces the field data acquisition scheme of 384 radar which is coherent and narrow-band. It introduces the calculation of the coherence error, direct shift, interference and noise of the system. The analyzing results of field data are given out.更多还原