测绘学报 ›› 2015, Vol. 44 ›› Issue (2): 183-189.doi: 10.11947/j.AGCS.2015.20130737

• 摄影测量学与遥感 • 上一篇    下一篇

多视角三维激光点云全局优化整体配准算法

李彩林1, 郭宝云1, 季铮2   

  1. 1. 山东理工大学 建筑工程学院, 山东 淄博 255049;
    2. 武汉大学 遥感信息工程学院, 湖北 武汉 430079
  • 收稿日期:2013-12-05 修回日期:2014-10-20 出版日期:2015-02-20 发布日期:2015-02-14
  • 通讯作者: 郭宝云E-mail:guobaoyun@sdut.edu.cn E-mail:guobaoyun@sdut.edu.cn
  • 作者简介:李彩林(1985—),男,博士,讲师,研究方向为数字摄影测量与计算机视觉、三维激光扫描数据处理等。E-mail:licailin@whu.edu.cn
  • 基金资助:

    国家自然科学基金(41301518); 四川省地理国情监测工程技术研究中心资助项目(GC201512);山东理工大学青年教师发展支持计划经费(114016);山东理工大学博士科研启动经费(413042; 413050)

Global Optimization and Whole Registration Algorithm of Multi-view 3D Laser Point Cloud

LI Cailin1, GUO Baoyun1, JI Zheng2   

  1. 1. Institute of Architecture and Engineering, Shandong University of Technology, Zibo 255049, China;
    2. School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China
  • Received:2013-12-05 Revised:2014-10-20 Online:2015-02-20 Published:2015-02-14
  • Supported by:

    The National Natural Science Foundation of China (No. 41301518);Geographic National Condition Monitoring Engineering Research Center of Sichuan Province (No. GC201512);The Young Teacher Development Support Foundation of Shandong University of Technology(No.114016);Doctoral Scientific Research Foundation of Shandong University of Technology (Nos. 413042;413050)

摘要:

提出一种已知多视激光点云配准初值进行自动全局优化的整体配准算法, 并详细推导了多视激光点云配准全局优化平差模型。本算法对多视角三维激光点云的扫描顺序不作要求, 可以处理无序散乱的多视三维激光扫描点云, 同时可以获得最小二乘意义下的最优变换参数, 实现多视三维激光点云的自动精确配准。利用实际三维激光扫描点云数据进行试验, 得到了预期的结果, 验证了本文方法的可行性和有效性。

关键词: 多视三维激光点云, 全局优化, 点云配准, 最小二乘平差, 整体配准模型

Abstract:

A global optimization and whole registration algorithm of multi-view 3D laser point cloud is presented. Detailed derivation of global optimization adjustment model of multi-view laser point cloud is showed in this paper. This algorithm can handle disordered and scattered multi-view 3D laser point cloud, at the same time optimal transformation parameters can be obtained. Practical 3D laser point cloud data are exemplified for the feasibility and effectiveness of proposed methods.

Key words: multiple-view 3D laser point cloud, global optimization, point cloud registration, least squares adjustment, whole registration model

中图分类号: