大规模光学遥感影像区域网平差方法研究

项目名称： 大规模光学遥感影像区域网平差方法研究

项目编号： No.40871167

项目类型： 面上项目

立项/批准年度： 2009

项目学科： 金属学与金属工艺

项目作者： 张力

作者单位： 中国测绘科学研究院

项目金额： 40万元

中文摘要： 目前，航空影像的空中三角测量物方坐标系是建立在地球切平面上的，数据处理范围受地理范围的限制。基于严密模型的航天遥感影像的处理主要还集中在单张影像或立体像对的定位定向上。随着高分辨率影像测图技术发展，通过区域网平差来实现稀少控制的大范围影像测图显得日益迫切。本课题拟在地心直角系下建立一套简洁的适合航空航天光学影像区域网平差模型，分析影像稀少控制条件下定位误差的存在规律，通过研究连接点误差方程式坐标未知数的消元、法方程式的改化以及大型稀疏矩阵分解求解方法，实现光学影像大规模区域网平差解算。通过实验分析，得到大范围光学影像区域网平差的重要性质，实现平差方法的优化，形成大范围影像区域网平差的理论、技术和方法。本项目的实施，将解决困难地区、境外地区以及月球光学影像大范围测图控制点稀少的难题。

中文关键词： 摄影测量;影像定位;空中三角测量;区域网平差;精度分析

英文摘要： Currently, the aerotriangulation object space coordinate system of aerial image is built on the tangent plane of the earth, and the scope of data processing is limited by the geographical scope. Aerospace remote sensing image processing based on rigorous model mainly focused on the position and orientation of single image or stereopair. With the development of high-resolution image mapping technology, achieve the image mapping of large area with few controls through block adjustment is becoming increasingly urgent. Based on geocentric rectangular coordinate system, this project aims to create a simple block adjustment model that fit aerospace optical image, analyzes laws of orientation error on condition that images have few controls, realizes the solution of large-scale block adjustment of optical image by studying the elimination method of tie point error equation, the reduction of normal equation,and the solving of large sparse matrix. Through experimental analysis, gains the important properties of large-scale block adjustment of optical image, optimizes the adjustment method, forms theories, technologies and methods of large-scale images block adjustment. In depressed regions, overseas areas and the moon, the project will be implemented to solve the difficulty of mapping large range areas with few control points.

英文关键词： photogrammetry;image positioning;block adjustment