| 汽车车身曲面点云数据分割方法的研究 |
| |
| 引用本文： | 王峰,王玉林,李旭.汽车车身曲面点云数据分割方法的研究[J].山东理工大学学报,2006,20(6):44-47. |
| |
| 作者姓名： | 王峰 王玉林 李旭 |
| |
| 作者单位： | 山东理工大学交通与车辆工程学院 山东淄博255049(王峰，李旭)，青岛大学机电工程学院 山东青岛266071(王玉林) |
| |
| 摘 要： | 车身曲面是由多张曲面经过延伸、过度、裁剪拼接而成,直接对测量获得的车身密集点云数据进行曲面拟合非常困难,需要对点云数据进行区域分割.探讨了点云数据的常用分割方法,结合车身曲面特点,研究并实现了一种点云分割的算法-基于平面度的直接分割方法.该方法通过最小二乘拟合平面法矢量夹角的均方差值来搜索具有几何相似特性的连续曲面片,将不同性质的曲面片分块保存.并结合ATOS测量设备得到的车身曲面点云数据,给出不同的实例,证明了该方法的有效性.
|
| 关 键 词： | 逆向工程 车身曲面 点云数据 数据分割 平面度 |
| 文章编号： | 1672-6197(2006)06-0044-04 |
| 收稿时间： | 03 14 2006 12:00AM |
| 修稿时间： | 2006年3月14日 |
Research in segmentation methods for point cloud of autobody surface |
| |
| WANG Feng,WANG Yu-lin,LI Xu.Research in segmentation methods for point cloud of autobody surface[J].Journal of Shandong University of Technology:Science and Technology,2006,20(6):44-47. |
| |
| Authors: | WANG Feng WANG Yu-lin LI Xu |
| |
| Institution: | 1. School of Traffic and Vehicle Engineering, Shandong University of Technology, Zibo 255049, China; 2. College of Mechanical and Electronic Engineering, Qingdao University, Qingdao 266071, China |
| |
| Abstract: | The autobody surface is formed by many curved surfaces through extending and cutting operations.Direct fitting surface with point cloud data obtained from autobody scanning is difficult.So the region segmentation of point cloud data is important.The common segmentation methods of cloud data were discussed in the paper.Considering the characteristics of autobody surface,a method for region segmentation algorithm according to the quantity planarity was proposed.This method uses the average squared error of the least squares plane fit to search continual curved surface piece which has the geometry similar characteristic,and preserves heterogeneity curved surface piece.This method has been validated through different examples,using the autobody point cloud data obtained from the ATOS measurement equipment. |
| |
| Keywords: | reverse engineering autobody surface point cloud region segmentation planarity |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
