无人机光束法空三稀疏矩阵处理与实现  

Adjustment of Bundle Aerotriangulation for UAV Images

作　　者：亓晨[1] 赵西安[1] 董友强 陈柯[2] 

机构地区：[1]北京建筑大学测绘与城市空间信息学院,北京100044 [2]北京灵图软件技术有限公司,北京100193

出　　处：《城市勘测》2014年第2期108-112,共5页Urban Geotechnical Investigation & Surveying

基　　金：国家自然科学基金(40771178);北京建筑大学产学研项目(csy2012026)

摘　　要：基于无人机影像的光束法空三是目前航空摄影测量领域的研究热点之一。本文简要分析了无人机航测与传统航测相比出现的技术难点;通过对无人机光束法空三平差中稀疏矩阵的分块存储和解算,节省了计算机内存的消耗,提高了运算速度,解决了无人机航测大数据量导致的问题;并通过对可能引起矩阵相关性的点的检测和剔除,有效避免了病态矩阵的出现,解决了无人机航测数据质量不好导致的问题。采用VS2010 C++开发实现,通过对无人机实验数据的处理,表明该算法具有可行性与高效性,满足1∶2 000的无人机测图要求。Bundle aerotriangulation based on UAV images is a hot topic in the research of aerial photogrammetry . This article briefly analysed UAV aerial technical difficulties compared with traditional aerial .By decomposing and cal-culating sparse matrixes existing in bundle aerotriangulation adjustment ,much computer memory is saved ,which acceler-ates operation and solves problems caused by large amount of data in UAV aerial ;Ill-conditioned matrix is avoided through detecting and removing points that may cause matrix correlation ,which solves problems caused by bad data in UAV aerial.The adjustment method has been developed in C＋＋.The result of experiment based on UAV images showed that the method is efficient and feasible .The adjusting results also may meet surveying and map in 1 ∶2000 .

关 键 词：无人机 光束法平差 稀疏矩阵 

分 类 号：P231[天文地球—摄影测量与遥感]