计量学报2024，Vol.45Issue(3)：433-439,7.DOI:10.3969/j.issn.1000-1158.2024.03.18
基于激光扫描仪的点云配准方法
Point Cloud Registration Method Based on Laser Scanner
摘要
Abstract
Aiming at the difficult problem of aligning the incomplete point cloud obtained from the actual scanning of the laser scanner,a registration method based on corresponding point pairs is proposed.Firstly,experiments are conducted by a laser scanner to obtain the measured point cloud data of the measured workpiece and configuration of Point Cloud Library environment based on Visual Studio software to investigate the point cloud alignment of real and ideal models.Secondly,the measured point cloud data are preprocessed by voxel filtering and uniform downsampling.Then,the point cloud data are aligned by corresponding point pairs to provide a better initial value of the transformation for the subsequent fine alignment.The point cloud alignment is realized based on the ICP algorithm to realize the fine alignment.Finally,the root-mean-square error is used as an evaluation index of the accuracy of the point cloud alignment to evaluate the results of the alignment,and the results are analyzed with the aid of the CloudCompare software.The visual display and analysis show that the root mean square error of the alignment can be controlled at 0.62 mm in the case where the measured workpiece itself is not absolutely smooth,which proves that the method has better results for the alignment of incomplete point clouds.关键词
几何量计量/激光扫描仪/点云配准/ICP算法/均方根误差Key words
geometric metrology/laser scanner/point cloud registration/ICP method/root mean square error分类
通用工业技术引用本文复制引用
褚海漫,程银宝,李亚茹,罗哉,江文松,王燕..基于激光扫描仪的点云配准方法[J].计量学报,2024,45(3):433-439,7.基金项目
浙江省自然科学基金(LY22E050005) （LY22E050005）
国家市场监管总局科技计划(2021MK188) （2021MK188）
