【作者】 杨杰；

【导师】 李德仁；

【作者基本信息】 武汉大学 ， 摄影测量与遥感， 2004， 博士

【摘要】 对遥感图像的应用，目标位置的确定是一件非常重要的事情。目标的位置参数在空间遥感中是一个基本的信息量，它是地形测绘、灾害监测、资源普查、变化检测等一切遥感应用的基础信息。如何对遥感图像进行精确定位及从遥感图像获取地面三维信息一直是遥感研究的重点和热点。合成孔径雷达(SAR)具有全天候、大范围、有一定穿透力等优点，被选作本文空间定位研究的对象。在SAR基础上发展的干涉合成孔径雷达(InSAR)，除继承了SAR的优点外，还能提取高精度地面三维的信息，能对地面目标的慢速运动和地壳微小移动做高精度的检测，其在军事、国民经济和科学研究中，有着极广的应用领域。对InSAR技术的研究，在国内外也受到了高度的重视。 SAR影像的象素定位方法通常有两种。一种做法是：在SAR影像中找出一些位置已经确定的参考点，然后推导影像象元相对于这些点的位置，从而实现SAR影像的象素定位；另外一种是无地面控制点的SAR影像象素定位方法，它的定位模型是Curlander J．C．提出的，并且运用Seasat卫星数据进行了实验，实现了用于海洋冰面运动的监测，精度可达到200米，但没有给出该模型的任何实用的求解方法。与第一种方法相比，该方法有如下优点：(1)不需要在星载SAR的视场中使用任何位置确知的参考点，因此将真正突破现存的地-空定位理论，达到空-空定位水平(绝对定位)；(2)其定位精度将仅仅取决于星历数据的准确性、地球模型的有效性、脉冲时延所决定的目标斜距精度、以及成像中多普勒信息的估计精度。因此对于SAR卫星，必须从成像几何特性去着手来实现无控制点的SAR影像精确定位。本文通过仿真研究，找出了各参数对定位精度的影响；通过真实数据的实验，证明了从模型入手的无地面控制点直接定位能获得较好精度。 InSAR的概念由L．C．Graham于1974年首先提出。由于其对雷达技术和成像技术要求较高，加之也没有专门发射用于干涉雷达处理的卫星，很难得到好的干涉数据，故在此后的一段时间发展比较缓慢。九十年代后，尤其是ERS-2卫星发射后，得到了大量的可用干涉雷达数据，极大地推动了InSAR发展，使其成为雷达和遥感技术研究领域最活跃的热点之一。 本文对SAR的研究，以星载数据为对象，对无地面控制点的数据实现直接定位，获取其定位精度，并对成像参数对定位的影响进行分析。对InSAR的研究，以最终获取数字高程图为目标。本文研究的主要内容有：成像的原理、成像算法对位置的影响、定位模型、InSAR的原理、InSAR完整的数据处理流程、流程中关键环节的理论和具体实现方法。对InSAR处理中各环节的论述，以理论为基础，力求简单准确，并给出切实可行的实现方法和准则.在对传统方法的分析基础上，对亚象素的配准，平地效应去除，相位的展开及相位到高程的转换等算法都进行了分析和改进，提出了自己的想法。最终真实数据的实验结果证明文中的分析和提出的算法是可靠的。 最后对本文所做工作进行了总结，提出了下一步在SAR和InSAR数据处理所应关注的目标，文章的基本章节顺序安排如下: 第一章叙述 SAR技术的现状和发展，提出了本文研究的主要内容; 第二章介绍了合成孔径雷达的基本原理及其主要工作方式; 第三章首先对星载SAR的儿何特点进行了分析;然后分析了SAR成像的运动补偿技术，并对星载SAR成像算法进行了分析介绍，最后实现了典型RD成像算法; 第四章详细讲述了SAR定位原理，建立了雷达成像参数对定位影响的关系曲线，并对点目标进行模拟研究来验证算法，得到了有价值的结论;最后应用实际数据对算法对模型进行了检验; 第五章阐述了InSAR的基本原理，介绍了对基线估计的方法，给出了影响InsAR高程测量精度的主要因素; 第六章给出了InSAR处理实现的整个流程，对各关键处理步骤进行了详细分析，并给出了关键部分的处理算法。最后在VC++环境下开发了InsAR处理软件，并结合实际数据进行了验证，表明整个处理流程的分析和处理是正确的、有效的; 第七章介绍了Gcolmager4.0的基本功能模块，并给出了InSAR算件的整个实现流程; 最后对本文所做工作进行了总结，提出了下一步在SAR和InSAR数据处理所应关注的目标。 综合本文的研究，取得的主要创新性成果概括如下: 1、分析成像算法的基础上，实现了星载SAR的R一D成像算法; 2、对5 AR的特性进行详细分析的基础上，完成了对点目标的仿真; 3、建立仿真模型，应用仿真模型对各项误差对定位精度的影响进行了理论分析。得出从成像来研究定位，多普勒中心频率对定位精度没有影响的结论: 4、运用实际数据对定位算法进行了实际验证，结果证明定位算法完全正确; 5、对inSAR数据处理进行了全面和系统的理论研究，并用VC++实现了InSAR的处理软件包; 6、在对传统方法进行分析的基础上，提出了一些新的方法和见解。更多还原

【Abstract】 How to determine location of target is very important in the application of remote sensing imagery. Location of target is an basic information in remote sensing imagery. It is fundamental for topographic survey, disaster monitoring, resource survey, change detection and so on. How to determine precise location and acquire 3D information by remote sensing is a focus in the remote sensing community. Synthetic Aperture Radar (SAR) is a hot point of research because SAR can observe all-weather and all day, and penetrate some object. Interferometric Synthetic Aperture Radar (InSAR) develops from SAR. Carried forward the merit of SAR, InSAR can acquire high precise DEM or accomplish change detection, such as slow-moving targets or subtle shift of the earth’s surface. InSAR is widely applicated in many areas, such as military, civil and scientific research, which make InSAR one of the most active fields in Radar and Remote Sensing nowadays.There are two methods of how to determine location of points in SAR imagery. First method: This method looks for some feature points whose precise location, then finish location of points in SAR imagery by deduce relative position of image pixel and the control points. Second method: Its model was first proposed by Curlander J.C in 1982. Tests were conducted using SEASAT SAR imagery, comparing predicted feature location with the location as determined by high precision area maps. Results indicate that location accuracy of 200 meters is attainable with this model. But he didn’t explain how to resolve the model in practical application. The method has two merits: 1) It needn’t determine location pixel in SAR imagery. The method will make a break in the theory of earth-space location; achieve space-space location (absolute location). 2) Accuracy of location will only be determined by precise of satellite ephemeris, earth model, precise of range by determined pulse delay, estimation of Doppler frequency. So we must study SAR imaging feature for implement location of SAR imagery with no ground control point. The test results are also promising.The concept of InSAR was first proposed by L. C. Graham in 1974. Because of its severe requirement on the raw data pair, no distinct progress was made in about ten years that followed. It is the huge amount of raw data achieved by various spacebome SAR systems (ERS, Radarsat,JERS) after 1990’s, which satisfy the request that make research on InSAR technique a focus in the remote sensing.In the paper, the location of target from SAR imagery is implemented without GCP and factors that effect position accuracy are analyzed. How to acquire precise DEM from InSAR is also studied in this paper. Main aspects of the paper include principle of SAR imaging, SAR imaging how to influence location of spaceborne SAR, model of location, principle of InSAR, implement chart of InSAR data processing, key principle and the fulfilling of each links of the procedure. Special stress is laid on the illustration to make basic concept and principles accurate, clear and easy to understand. After analyzing traditional algorithm, how to improve accurate of subpixel register, flat-earth remove, phase unwrapping and phase transfer to height is brought forward. Results indicate that algorithm and model is reasonable. Arrangement of chapters and sections are as follows:Chapter One presents development and current state of SAR, bring forward main research contents.Chapter Two discuss basic principle and primary working of spaceborne SAR.Chapter Three analyses geometry of SAR, motion compensation of SAR, and algorithm for spaceborne SAR imaging, implement RD algorithm.Chapter Four discusses principle of location of spaceborne SAR, describes the relation between imaging parameter and location of object. Implement simulation of point object and acquire an important conclusion. Final tests were conducted using real data.Chapter Five presents basic principle of InSAR, introduce algorithm for estimate baseline. Primary factors which influence the accuracy of DEM obtained f更多还原