大型结构件焊接型面自动检测技术研究  

Research on Automatic Detection Technology of Welding Profile of Large Structural Parts

作　　者：汪林 曹春平[1] WANG Lin;CAO Chunping(School of Mechanical Engineering,Nanjing University of Science and Technology,Jiangsu Nanjing 210094,China)

机构地区：[1]南京理工大学机械工程学院,江苏南京210094

出　　处：《机械设计与制造》2024年第12期237-242,共6页Machinery Design & Manufacture

基　　金：江苏省重点研发项目—高效精密热模锻生产系统智能控制技术研究(BE2015011-3)。

摘　　要：大型结构件的焊接型面质量检测在成形加工前至关重要,随着现代工业生产规模的扩大,依赖人工测量并与图纸比较的传统检测方式正在变得不可行,现有的三维扫描检测也因耗时过长而无法使用,需要一种更加高效和稳健的方法来支撑工厂的检测。为此,这里提出了一种基于特征匹配的大型结构件焊接型面检测方法。首先,利用三维扫描仪采集型面焊接特征的点云数据,对数据进行K邻域去噪、体素下采样的处理操作;其次,采用先坐标粗配准再ICP算法精配准的方式对齐点云数据与CAD模型,再通过基于距离的偏差分析和几何参数比较评估特征型面质量;最后,以某大型压力机焊接型面为例,通过对距离偏差和装配孔方位偏差的检测,验证该方法不仅满足检测精度的要求,而且效率提高了近三倍。The welding profile quality inspection of large structural parts is very important before forming and processing.With the expansion of modern industrial production scale,traditional inspection methods that rely on manual measurement and comparison with drawings are becoming infeasible,the existing 3D scanning inspection is also unusable because it takes too long,a more efficient and robust method is needed to support the factory inspection.For this reason,this paper proposes a welding profile detection method for large-scale structural parts based on feature matching.Firstly,the 3D scanner was used to collect the point cloud data of welding features,the K-neighborhood denoising and voxel downsampling were performed on the data;Then,the point cloud data and the CAD model were aligned with the coordinate coarse registration first and then the ICP algorithm fine registration,the surface quality was evaluated by distance based deviation analysis and geometric parameter comparison;Finally,taking the welding surface of a large press as an example,through the detection of distance deviation and assembly hole orientation deviation,it is verified that the method not only meets the requirements of detection accuracy,but also improves the efficiency by nearly three times.

关 键 词：特征匹配 焊接型面 三维检测 点云处理 质量评估 

分 类 号：TH16[机械工程—机械制造及自动化] TP391.41[自动化与计算机技术—计算机应用技术]