应用科学学报 ›› 2015, Vol. 33 ›› Issue (4): 399-406.doi: 10.3969/j.issn.0255-8297.2015.04.006

• 信号与信息处理 • 上一篇    下一篇

一种用于全景影像的测量方案

孙振兴1,2,3, 钟若飞1,2,3   

  1. 1. 首都师范大学资源环境与地理信息系统北京市重点实验室,北京100048
    2. 首都师范大学城市环境过程与数字模拟国家重点实验室培育基地,北京100048
    3. 首都师范大学三维信息获取与应用教育部重点实验室,北京100048
  • 收稿日期:2014-12-11 修回日期:2015-04-06 出版日期:2015-07-30 发布日期:2015-04-06
  • 作者简介:钟若飞,教授,博导,研究方向:车载激光测量,E-mail: zrfsss@163.com
  • 基金资助:

    国家自自然科学基金(No.41371434, No.41261086);测绘地理信息公益性行业科研专项经费项目基金(No.201412020)资助

Measurement Scheme for Panoramic Images

SUN Zhen-xing1,2,3, ZHONG Ruo-fei1,2   

  1. 1. State Key Laboratory Incubation Base of Urban Environmental Processes and
    Digital Simulation, Capital Normal University, Beijing 100048, China
    2. Beijing Key Laboratory of Resource Environment and Geographic Information System,
    Capital Normal University, Beijing 100048, China
    3. Key Laboratory of 3D Information Acquisition and Application, Ministry of Education,
    Capital Normal University, Beijing 100048, China
  • Received:2014-12-11 Revised:2015-04-06 Online:2015-07-30 Published:2015-04-06

摘要: 提出了一种从全景影像中还原目标三维信息的测量方案. 通过全景相机所获取的影像只有GPS 坐标,而没有姿态信息,需要运用影像匹配技术以及光束法准确算出拍摄时刻全景影像的姿态,然后通过前方交会计算全景影像中目标的三维坐标. 针对全景影像畸变较大的特点,采用仿射不变特征匹配算法进行影像匹配,同时使用随机抽样一致算法剔除粗差点,以保证匹配点的数量及准确度. 根据全景影像的透视投影几何模型,能改进常规摄影测量中的光束法,可用来求解全景影像姿态. 提出一种针对全景摄影测量的前方交会算法,将空间直线方程变换后建立法方程,并进行平差解算. 实验表明,该方法相比于传统方法有更高的精度,可准确测算全景影像上的物点坐标.

关键词: 全景测量, 影像匹配, 光束法平差, 前方交会

Abstract: This paper presents a measurement solution that recovers 3D information of
objects from a panoramic image. In case of having GPS position only, the panoramic
posture at the shooting time can be calculated by image match and bundle adjustment.
Thus we can use intersection to obtain the objects’ coordinates. As large distortion in
panoramic is an obstacle in image match, an affine invariant feature matching algorithm
is used for image match, and a random sample consensus algorithm used to eliminate
coarse errors, giving better point match in terms of both quantity and quality. For the unique geometric feature of panoramic image, the conventional photogrammetric bundle
adjustment is improved and used to solve posture of panorama. Finally, an intersection
method is proposed to build normal equations through transformation of space straight
line equations before adjustment. Compared with the conventional method, the proposed
algorithm has higher accuracy. Experimental results show feasibility of the scheme. It can
accurately measures object coordinates in panoramic images.

Key words: panoramic measurement, image match, bundle adjustment, photogrammetric triangulation

中图分类号: