【作者】 李贺；

【导师】 秦志远；

【作者基本信息】 解放军信息工程大学 ， 模式识别与智能控制， 2009， 硕士

【摘要】 合成孔径雷达(SAR, Synthetic Aperture Radar)因具有全天时、全天候以及一定的穿透能力等优点,在世界范围内受到极大重视并获得飞跃发展。SAR图像模拟是根据地面实际情况或其它遥感资料等数据,用计算机产生一幅SAR图像,以供多种应用的技术。随着SAR的发展,SAR图像模拟技术也在不断进步,并已在国防和经济建设等多个领域发挥着重要的作用,如:用于雷达系统设计与验证、雷达图像几何精校正和地理编码、雷达图像解译和目标识别、测试新算法等方面。但在国内SAR图像模拟及其应用的理论和技术还不够成熟,因此研究SAR图像模拟具有紧迫性和实用性。SAR图像模拟方法分为基于原始信号的模拟方法和基于图像特征的模拟方法,本文重点研究了基于图像特征的模拟方法。详细阐述了基于图像特征的SAR图像模拟的原理、实现方法和应用,设计了不同的SAR图像模拟方案,并进行了仿真实验。本文的主要内容和工作如下:1、总结分析了SAR图像模拟的各种方法,简要介绍了各种模拟方法的主要原理及步骤。给出了SAR图像的几何特性及SAR图像固有的斑点噪声模型。2、详细介绍了基于图像特征的SAR图像模拟所涉及的理论。介绍了DEM数据、轨道数据和坐标转换等数据预处理方法,分析了常用的构像模型和地物后向散射模型。3、设计并实现了两种情况下的SAR图像模拟方法。在地形起伏比较明显的山区采用基于DEM的SAR图像模拟方法;在平地等地物类别较多的地方采用基于地物分类信息和DEM的SAR图像模拟方法。4、在强度模拟时,根据SAR成像的实际情况,采用基于强度积分的策略,丰富了模拟图像灰度级,使图像更细腻。5、实现了斑点噪声模拟,根据乘性噪声模型对光滑的SAR模拟图像加入斑点噪声,改善了图像的视觉效果。在InSAR方面做了尝试,实现了InSAR复图像的模拟和干涉图的模拟。6、利用模拟图像实现了SAR图像的几何校正。利用已知的DEM生成模拟SAR图像,然后利用一种尺度不变特征转换算子--SIFT算子对真实图像和模拟图像进行配准,通过模拟图像建立起DEM与真实图像之间的变换关系,从而根据构像模型实现SAR图像的几何精校正。更多还原

【Abstract】 With the advantage of working day-and-night, of all-weather and the ability of penetrating something, SAR has been attached great importance all over the world and developed fleetly. SAR image simulation is a technology of producing a SAR image with computer for multiplicate applications according to ground data or other remote sensing data etc. The technology of SAR image simulation was developing rapidly with the development of SAR, and had played an important role in national defence and economic fields, e.g. the design and validation of radar system, the precise geometrical rectification and geocoding of radar images, the interpretation and recognition of radar images, testing new arithmetic etc. But the theory and technology of the simulation of SAR image and its application has not been developed deeply. So it is necessitous and practical to study the technology of SAR image simulation.SAR image simulation technology is classified into two methods, one is the simulation method based on raw signal, and the other is simulation method based on image characteristics. The thesis studied mostly on the method based on image characteristics. The simulation theory, the method of carrying out simulation and the application of simulation in this thesis were all included. Two means of simulation in this thesis were designed according different cases and the experiments were done at last.The central content and work in this thesis were followed as below:1. Summarizing and analyzing all kinds of SAR image simulation methods,giving a brief introduction of the theories and processes of all the methods. The geometric characteristics and the model of inherent speckle noise were given.2. Introducing the knowledge referring to SAR image simulation based on image characteristic particularly. The common methods of data preprocessing of DEM data, orbital data and the transformation of coordinates were given. And the imagery models and backscattering models of objects on the land were analyzed.3. The simulation methods of SAR image at two cases were designed and carried out. When there was obvious hypsography the simulation method based on DEM was adopted; When the land was flat and there were sorts of land covers, the simulation method based on DEM with classification information of land objects was adopted.4. The strategy of intension integral was adopted when computing the intension of SAR image according to the imagery fact. And with this method the hue was improved and the detail was enriched.5. The simulation of speckle noise was realized. The reality by adding speckle noise into the smoothing image according to the multiplied noise model was improved. Simulation of InSAR complex image and simulation of interferogram were implemented as the attempt in InSAR. 6. Realizing the geometric rectification of SAR image by using the simulation SAR image. Firstly, simulating SAR image through given DEM data, then matching the simulation image and the real image with SIFT arithmetic operator so as to build the transforming relation between real image and DEM coordinate, and then the geometric rectification of SAR image could be implemented with the imagery model.更多还原