基于点云三维重建的装配件尺寸测量系统  

Dimensional Measurement System of Assembly Parts Based on 3D Reconstruction of Point Cloud

作　　者：余永维 徐维 YU Yong-wei;XU Wei(School of Mechanical Engineering,Chongqing University of Technology,Chongqing 400054,China)

机构地区：[1]重庆理工大学机械工程学院,重庆400054

出　　处：《组合机床与自动化加工技术》2022年第6期105-109,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：重庆理工大学研究生科研创新项目(clgycx20202082,clgycx20202073);重庆市专业学位研究生教学案例库项目(2019-79)。

摘　　要：针对机械加工偏差和人为因素,造成装配面间几何尺寸信息不匹配,而采用人工检测效率低、重复性差,难以满足工厂在线实时检测要求等问题,提出一种基于点云三维重建的装配件实时在线测量系统。以点与点间存在的内在约束作为先验引导信息,提出改进的RANSAC算法来检测并剔除异常数据点,显著减少算法耗时同时大幅提高了平面特征的提取准确率;再利用最小二乘法对剩余有效点进行拟合,对法兰盘工件重建三维模型并进行关键尺寸测量。实验结果表明,基于点云三维重建的工件尺寸测量方法有较好的准确性和检测效率,法兰盘工件测量过程耗时约5.6 s,测量标准误差在±0.16 mm以内,测量误差不超过0.30 mm,满足了工件测量的实时性和准确性要求。In order to solve the problems of the mismatch of geometric dimension information between assembly surfaces caused by machining deviation and human factors,and the low efficiency and poor repeatability of manual detection,it is difficult to meet the requirements of factory online real-time detection,A real-time online measurement system of assembly parts based on point cloud 3D reconstruction is proposed.An improved RANSAC algorithm is proposed to detect and eliminate abnormal data points by using the inherent constraints between points as prior guidance information,which significantly reduces the time consuming of the algorithm and improves the extraction accuracy of plane features.Then the least square method is used to fit the remaining effective points,and the three-dimensional model of the flange workpiece is reconstructed and the key dimensions are measured.The experimental results show that the workpiece size measurement method based on point cloud three-dimensional reconstruction in this paper has good accuracy and detection efficiency.The measurement process of flange workpiece takes about 5.6 s,the standard measurement error is less than±0.16 mm,and the measurement error is less than 0.30 mm,which meets the real-time and accuracy requirements of the workpiece measurement.

关 键 词：点云数据 视觉测量 RANSAC 最小二乘法 平面特征 

分 类 号：TH161[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]