车身漆面缺陷的自动检测与定位方法  

Automatic Detection and Location Method of Car Body Paint Defects

作　　者：蒋良犇 陈正涛 沈彬[1,2] 王皓[1,2] JIANG Liang-ben;CHEN Zheng-tao;SHEN Bin;WANG Hao(School of Mechanical Engineering,Shanghai Jiaotong University,Shanghai 200240,China;Fraunhofer Project Center for Smart Manufacturing at Shanghai Jiaotong University,Shanghai 201306,China)

机构地区：[1]上海交通大学机械与动力工程学院,上海200240 [2]上海交通大学弗劳恩霍夫协会智能制造项目中心,上海201306

出　　处：《机械设计与制造》2024年第10期139-144,共6页Machinery Design & Manufacture

基　　金：弗劳恩霍夫-上海交通大学智能制造项目(ZN2017020102)。

摘　　要：针对目前国内汽车制造业中面漆后车身表面缺陷难以非接触高精度检测,得不到缺陷确切位置等问题,这里提出了一种结合相位测量偏折术与双目测距原理的车身漆面缺陷检测定位方法。首先采用四步相移法提取车身反射图像相位,运用格雷码进行相位解包裹,再通过判断局部相位突变来提取缺陷;建立缺陷池并通过过滤器去除伪缺陷得到真实缺陷;基于双目测距原理,对检测出的缺陷,取其形状中心再得到具体位置坐标。实验结果表明,该缺陷检测方法能够自动,准确,稳定地检测漆面缺陷;双目测距法能准确地得到缺陷的具体位置,可为后续实现自动打磨提供可靠数据来源。Aiming at the current domestic automobile manufacturing industry’s difficulty in non-contact and high-precision detection of surface defects after top coating,and the inability to obtain the exact position of the defect,this paper proposes a car body painted surface defect detection and location method that combines the phase measuring deflectometry and the principle of binocular passive ranging.First,the four-step phase shift method is used to extract the phase of the body reflection image,and the Gray code is used for phase unwrapping,and then the defects are extracted by judging the local phase mutation;establish a defect pool and remove false defects through filters to obtain real defects;based on the principle of binocular distance measurement,the shape center of the detected defect is taken to obtain the specific position coordinate.Experimental results show that the defect detection method can detect paint defects automatically,accurately and stably;the binocular distance measurement method can accurately obtain the specific location of the defect,which can provide a reliable data source for subsequent automatic polishing.

关 键 词：车身缺陷检测 相位测量偏折术 四步相移 双目测距 缺陷定位 

分 类 号：TH16[机械工程—机械制造及自动化] U462.1[机械工程—车辆工程] N94[交通运输工程—载运工具运用工程]