高技术通讯2017，Vol.27Issue(4)：351-358,8.DOI:10.3772/j.issn.1002-0470.2017.04.008
平面参数空间的实时三维点云配准方法
A method for real-time 3D point cloud registration in a parameter space of planes
摘要
Abstract
For 3D localization and modelling of an indoor environment with a multi-plane structure, a new method for real-time 3D point cloud registration in a planar parameter space was proposed.The method uses an improved 3D hough transform to quickly extract the plane features in point clouds, then, searches the closest planes in the parameter space of the planes by using an iterative method, and finally, estimates the translation and rotation of the point clouds by registration of the correspoding planes.A comparison experiment performed in a laboratory environment showed that the registration accuracy of the proposed method reached the traditional iterative closest point (ICP) method, and the registration speed was greatly improved.The new method can achieve the real-time registration of point clouds when using an ordinary laptop CPU.关键词
点云配准/平面提取/霍夫变换/三维地图重建/迭代最近点(ICP)Key words
point cloud registration/plane extraction/Hough transform/3D map reconstruction/iterative closest point (ICP)引用本文复制引用
王力宇,曹其新,王雯珊..平面参数空间的实时三维点云配准方法[J].高技术通讯,2017,27(4):351-358,8.基金项目
国家自然科学基金(61273331)资助项目. （61273331）
