摘要
在全景影像和激光点云数据的匹配中引入"虚拟投影球面"的假设,提出了基于光源归一的投影回归方法,并对该方法进行了严密的数学论证,在此基础上,采用球面投影中的角度逼近法进行了不同角度分辨率下全景影像的空间测量实验,实验结果验证了该方法的实用性。
Aimed to match panoramic images with laser point cloud data, we introduced an assumption of ＂virtual spherical projection＂ based on fisheye camera optical structure research and the mathemati cal basis of fisheye image. A projection regression method was proposed to normalize light sources for the fisheye camera lens and the laser scanner, and a mathematical processing rigorously supportes our hypothesis. We demonstrate the correctness of a angle approximation approach for spherical projection, and also measured panoramic images under different resolutions of scanner pith angle using nearest point search algorithm. Our experimental results show small measurement errors and there fore the proposed method has great practicality.
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2014年第1期32-36,共5页
Geomatics and Information Science of Wuhan University
基金
国家自然科学基金青年基金资助项目(41101449)
精密工程与工业测量国家测绘地理信息局重点实验室开放研究基金资助项目(PF2011-26)~~
关键词
全景影像
点云匹配
投影回归
角度逼近
测量
panoramic images
point cloud matching
projection regression
angle approximation ap- proach
measurement
