• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2018, Vol. 54 ›› Issue (15): 141-149.doi: 10.3901/JME.2018.15.141

• 数字化设计与制造 • 上一篇    下一篇

扫码分享

基于局部泊松曲面重建的点云刚性配准方法

孙殿柱, 郭洪帅, 李延瑞, 聂乐魁   

  1. 山东理工大学机械工程学院 淄博 255000
  • 收稿日期:2017-09-18 修回日期:2018-02-06 出版日期:2018-08-05 发布日期:2018-08-05
  • 通讯作者: 孙殿柱(通信作者),男,1956年出生,博士,教授,博士研究生导师。主要研究方向为逆向工程、数字化设计与制造。E-mail:sundianzhu@sdut.edu.cn
  • 作者简介:郭洪帅,男,1989年出生,硕士研究生。主要研究方向为逆向工程。E-mail:guohs96@163.com;李延瑞,男,1979年出生,博士研究生。主要研究方向为逆向工程。E-mail:liyanrui.m2@gmail.com;聂乐魁,男,1991年出生,硕士研究生。主要研究方向为逆向工程。E-mail:qingfengxiao135@163.com
  • 基金资助:
    国家自然科学基金(51575326)和山东省自然科学基金(ZR2015EM031)资助项目。

Method of Rigid Registration Based on Poisson Reconstruction of Local Sample Points

SUN Dianzhu, GUO Hongshuai, LI Yanrui, NIE Lekui   

  1. School of Mechanical Engineering, Shandong University of Technology, Zibo 255000
  • Received:2017-09-18 Revised:2018-02-06 Online:2018-08-05 Published:2018-08-05

摘要: 为兼顾点云配准精度与收敛速度,提高配准过程收敛的稳定性,提出一种基于局部泊松曲面重建的点云刚性配准方法。该方法基于曲面局部样本重建泊松曲面并将其作为配准参考曲面,以采样点至参考曲面的最近点作为匹配点,将匹配范围由曲面局部样本扩大为该样本所拟合的曲面域,提高点对匹配精度。同时,基于所重建的泊松曲面计算配准误差,消除采样误差对配准过程收敛判定的延迟影响,加快配准过程的收敛速度。在配准过程中,通过调整因子调节匹配点在曲面域的匹配范围,增大点对在配准误差偏大情况下匹配的概率,提高配准过程收敛的稳定性。试验表明,该方法在保证配准精度的同时可显著减少收敛时的迭代次数,配准过程收敛于全局最优的稳定性更高。

关键词: 泊松重建, 迭代最近点, 刚性配准, 曲面局部样本

Abstract: To keep a balance between the accuracy and the convergence rate of registration, and to improve the convergence stability of registration process, a method of rigid registration based on Poisson reconstruction of local surface sample is proposed. The method is based on the Poisson surface reconstructed by local surface sample and the surface can be regarded as the reference of registration. The nearest point from the point set in the local sample to the reference surface can be used as the match point, which can extend the match range from the point set to the surface fitted by the set. As a result, the accuracy of points matching can be improved. Registration error can be estimated by the Poisson surface, eliminating the delayed effects on convergence judgment caused by the sampling error and accelerating the speed of convergence of the registration process. In the registration process, adjustment factor is used to adjust the match range of the match point. The factor can increase the probability of matching in the case of a relatively large error and improve the convergence stability of registration process. Tests show that the method proposed can guarantee the accuracy of registration and significantly reduce the number of iterations when convergence reached at the same time. In addition, the method can reach global convergence more stably.

Key words: iterative closet point, local surface sample, Poisson reconstruction, rigid registration

中图分类号: