【作者】 贾桂敏；

【导师】 王向军；

【作者基本信息】 天津大学 ， 测试计量技术及仪器， 2013， 博士

【摘要】 推扫式三线阵CCD相机在星载和机载遥感立体测量中得到了广泛应用。由于其特殊的成像方式，遥感平台姿态变化、地面形貌变化、遥感平台振动、地球自转等因素对三线阵推扫影像及立体测量精度的影响更为复杂。本文针对姿态变化对遥感推扫式三线阵CCD相机影像像移及立体测量精度影响的关键理论和技术展开研究，涉及的研究内容如下：1.深入和系统地阐述了遥感推扫式线阵CCD相机的发展历程，典型单线阵、双线阵和三线阵CCD相机的成像参数和立体测量参数；研究了遥感推扫式三线阵CCD相机立体成像原理及其影像特征，立体摄影测量中涉及的常用像空间坐标系及其转换关系；综述了遥感推扫式三线阵CCD相机常用成像数学模型，以及各种模型的适用范围及定位精度。2.研究了姿态变化对遥感推扫式三线阵CCD相机成像像移的影响。利用坐标系转换方法建立了三线阵CCD相机的前向像移数学模型,构建了三轴姿态角变化对三线阵CCD相机成像影响原理图，提出了三轴姿态变化时遥感推扫式三线阵CCD相机成像像移数学模型。在此基础上，利用现有航空三线阵CCD相机参数，对三轴姿态变化时三线阵CCD相机不同线阵列的像移分布、不同姿态角对像移的影响程度及多姿态角并行变化时像移的变化规律进行了仿真分析，得到了三轴姿态变化对推扫式三线阵CCD相机成像的影响规律。3.提出了基于空间前方交会原理的无控制点遥感推扫式三线阵CCD相机立体摄影测量误差模型。利用Chang’E-1三线阵CCD相机成像参数、在轨运行相关参数、Chang’E-1三线阵CCD相机2C级影像数据，针对基线、姿态角、像点坐标、焦距对Chang’E-1三线阵CCD相机无控制点立体摄影测量精度的影响进行了实验和分析，对比了不同姿态角对Chang’E-1三线阵CCD相机立体定位精度的影响。4.提出一种半实物仿真实验方法来研究姿态变化对遥感推扫式三线阵立体测量精度的影响，完成了实验系统的搭建，实现了影像数据的采集和三线阵图像的提取；提出一种基于地形复杂度定量分析和地形分类的DEM建模方法选择策略，实现了DEM建模精度和建模速度的统一，为月面形貌实物模型的制备提供了理论基础；对三轴姿态变化时的标志点高程测量误差进行了计算和分析，验证了不同姿态角对高程定位精度的影响规律，并对连续滚转角变化时的三线阵图像进行了像移补偿，验证了提出的三线阵CCD相机像移数学计算模型。更多还原

【Abstract】 Three-line-array (TLA) push-broom cameras have been widely used in remotesensing stereo photogrammetry, eithor sapceborne or airborne. Due to the specialimaging mechanism, the influence that brought by many factors such as attitudevariation, topographic relief, vibration, rotation of earth, is more complicated thanother cameras. Key thoeries and technologies of attitude variation effects on TLApush-broom cameras in imaging and stereophotogrammetry are studied in this paper.The main research contents are as follows:1. The development of push-broom linear array CCD cameras has been reviewed,and many typical cameras of single linear array CCD, dual linear array CCD and threelinear arrays CCD are introduced in detail. Both imaging parameters andstereophotogrammetry parameters of these linear array CCD camers are listed.Stereoimaging principle and features of images of TLA push-broom cameras arestudied. The common coordinate systems of image space and their transformationrelation are intoduced. The common used mathematical models for TLA push-broomcameras and their application scope and positioning precision are also reviewed.2. The image motion model for TLA push-broom cameras is presented in this paper.Both aircraft velocity and attitude instability are taken into account in modeling imagemotion. Effects of aircraft pitch, roll, and yaw on image motion are analyzed based ongeometric relations in designated coordinate systems. Quantitative analysis to imagemotion is then conducted in simulation experiments. The results have shown thatimage motion caused by aircraft velocity is space invariant whilst image motioncaused by aircraft attitude instability is more complicated. Pitch, roll, and yaw allcontribute to image motion to different extents. Pitch dominates the along-track imagemotion and both roll and yaw greatly contribute to the cross-track image motion.These results provide a valuable base for image motion compensation to ensure highaccurate imagery in aerial photogrammetry.3. The positioning error model for stereophotogrammetry using TLA push-broomcameras is established base on space intersection principle. Experiments are carriedout based on Chang’E-1TLA push-broom camera parameters of imaging and orbitand attitude control. Effects of three attitudes (pitch, roll and yaw), baseline, stereoimage pair coordinates and focal length are analyzed. 4. An approach of semi-physical simulation for attitude instability using remotesensing TLA push-broom cameras is presented. The design scheme for three-axisattitude control module is described, and two physical models of high precision aremade using real lunar terrain elevation data downloaded from NASA. An effectiveapproach for selection of appropriate terrain modeling methods in forming a digitalelevation model (DEM) is proposed. This approach achieves a balance betweenmodeling accuracy and modeling speed. A support vector machine classifies terrainsurfaces into either complex or moderate based on a terrain complexity indexassociated with the terrain elevation range. The attitude variation is conducted insemi-physical simulation experiments, the original videos are captured, and imagestrips of TLA camera are extracted from these videos. Elevation errors of the fivemark points, located on the physical terrain models, which are produced by attitudesvariation, are calculated and analyzed, based on these image strips, and the effectrules on the elevation positioning accuracy caused by attitudes are proved. The imagemotion caused by rolling is compensated based on the mathematical models that areproposed in this paper, and the compensation result proves this model is correct.更多还原