基于机器视觉的汽车零部件平面度检测系统设计  

Research on flatness inspection system for automotive parts based on machine vision

作　　者：王宇龙 Wang Yulong(College of Metrology&Measurement Engineering,China Jiliang University,Hangzhou Zhejiang,310018)

机构地区：[1]中国计量大学计量测试工程学院,浙江杭州310018

出　　处：《电子测试》2022年第18期31-33,共3页Electronic Test

摘　　要：针对目前国内汽车零部件平面度检测存在的检测精度不高、检测速度较慢、部分检测过程仍采用人工检测等问题,本文设计了基于机器视觉的汽车零部件平面度在线检测系统。硬件部分以激光轮廓仪为基本组成部分;软件部分基于PCL点云库,采用C++算法编写可视化软件。采取先粗配准再精配准的点云配准方式。实验结果表明,本系统测量精度可达0.01mm,测量的总时间可控制在24.5s内,因此本系统适用于需要大批量检测汽车零部件平面度的场合。In view of the current problems of the inspection of flatness of domestic auto parts,such as low inspection accuracy,slow inspection speed,and part of the inspection process still using manual inspection,this paper designs an online inspection system for flatness of auto parts based on machine vision.The hardware part takes laser profiler as the basic component;the software part is based on PCL point cloud library and C++ algorithm is used to write the visualization software.The point cloud alignment method of coarse alignment and then fine alignment is adopted.The experimental results show that the measurement accuracy of this system can reach 0.01mm and the total measurement time can be controlled within 24.5s.Therefore,this system is suitable for applications that need to inspect the flatness of automotive parts in large quantities.

关 键 词：机器视觉 平面度测量 点云配准 ICP算法 SAC-IA FPFH特征描述 

分 类 号：TH821[机械工程—仪器科学与技术]