汽车覆盖件三角配合区域无标定尺寸测量方法  

Measurement Method for No Demarcated Triangle Fit Areas of Automotive Covering Components

作　　者：陆雨薇 蹇松阳 吕俊成 刘伯堃 LU Yu-wei;JIAN Song-yang;LYU Jun-cheng;LIU Bo-kun(School of Mechanical and Automotive Engineering,Guangxi University of Science and Technology,Liuzhou 545006;School of Mechanical Engineering,Shanghai Jiao Tong University,Shanghai 200240;SAIC-GM-Wuling Co.,Ltd.,Technical Center,Liuzhou 545007)

机构地区：[1]广西科技大学机械与汽车工程学院,广西柳州545006 [2]上海交通大学机械与动力工程学院,上海200240 [3]上汽通用五菱汽车股份有限公司广西新能源汽车实验室,广西柳州545007

出　　处：《制造业自动化》2025年第4期79-89,共11页Manufacturing Automation

基　　金：广西省重点研发计划(2023AB38002);广西省科技计划(桂科AA24206060);博士后专项基金(277291)。

摘　　要：针对整车自动化质检环节中汽车覆盖件三角配合区域尺寸测量问题,提出了一套基于图像的无标定尺寸测量方法,包括图像畸变矫正和尺寸测量两部分内容。首先,通过设计对照板校正切向畸变;其次,利用像素特征信息估计并矫正径向畸变,提高测量精度;最后设计了三角配合区域轮廓提取方法,确定最大内切圆尺寸,提高测量稳定性。实验结果表明,该方法矫正了一定程度的径向畸变和切向畸变误差,三角配合区域最大内切圆尺寸测量偏差在0.2 mm以内,成功率达到了98%。这一研究为整车自动化质检提供了高精度、高效率、简便的测量方法。In addressing the issue of measuring the dimensions of the triangular mating area of automobile body components in the context of automated quality inspection for the entire vehicle,a non-calibration-based measurement method using image processing has been proposed.This method comprises two main components:image distortion correction and dimension measurement.Firstly,tangential distortion is corrected by designing a reference board.Secondly,radial distortion is estimated and corrected using pixel feature information to enhance measurement accuracy.Finally,a method for extracting the contour of the triangular mating area is designed to determine the maximum inscribed circle size,thereby improving measurement stability.Experimental results demonstrate the successful correction of a certain degree of radial and tangential distortion errors.The measured maximum inscribed circle size deviation for the triangular mating area is within 0.2 mm,achieving a success rate of 98%.This research provides a high-precision,efficient,and straightforward measurement method for automated quality inspection of entire vehicles.

关 键 词：自动化检测 机器视觉 算法设计 畸变矫正 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]