金属薄片零件微小变形缺陷视觉检测方法与应用  

Metallic Thin Sheet Component Micro Deformation Defect Visual Inspection Method and Its Application

作　　者：孙成 陈冠浩 邓耀华[2] Sun Cheng;Chen Guanhao;Deng Yaohua(Vitalink Technology Co.,Ltd.,Xianning,Hubei 437300,China;School of Electromechanical Engineering,Guangdong University of Technology,Guangzhou 510006,China)

机构地区：[1]维达力科技股份有限公司,湖北咸宁437300 [2]广东工业大学机电工程学院,广州510006

出　　处：《机电工程技术》2024年第10期157-162,共6页Mechanical & Electrical Engineering Technology

基　　金：广东省省级科技计划项目(2023A0505050151);东莞市重点领域研发项(20221200300042);河源市科技计划项目(230511101473496)。

摘　　要：金属薄片零件常见于新能源汽车电池、精密电子产品装配中的端子连接件,其外形尺寸是否符合要求对成品组装质量至关重要。针对金属薄片零件存在微小变形,影响组装成品质量的问题,以金属薄片装配分析和缺陷检测要求为切入点,研究金属薄片零件微小变形缺陷视觉检测方法,在基于Canny边缘检测的基础上,引入改进的Devernay亚像素边缘技术,实现金属薄片零件边缘关键特征检测,结合并行处理方式进行边缘特征拟合,得到金属薄片零件变形评估数据,最后研发了金属薄片零件微小变形缺陷检测算法,有效提高金属薄片零件检测精度和检测速度。实验结果表明,所提算法的检测精度达到0.01mm,检测平均误差在0.03mm以内,检测平均时间为327ms,满足实际生产过程检测精度的要求。Metal thin sheet components are commonly found in terminal connectors used in the assembly of new energy vehicle batteries and precision electronic products.The conformity of their dimensional specifications is crucial for the product assembly quality.Addressing the issue of minor deformations in metal thin sheet components affecting the quality of the assembled product,the requirements of metal thin sheet assembly analysis and defect detection are taken as a starting point,visual detection method for micro deformations in metal thin sheet components is studied.Building upon the foundation of Canny edge detection,the study introduces an improved Devernay subpixel edge technology to detect crucial edge features of metal thin sheet components.By employing parallel processing for edge feature fitting,the deformation assessment data for metal thin sheet components is obtained.Finally,a novel algorithm for detecting micro deformations in metal thin sheet components is developed,effectively enhancing the detection accuracy and speed.Experimental results demonstrate that the algorithm achieves a detection accuracy of 0.01 mm,with an average error within 0.03 mm and an average detection time of 327 ms.The algorithm meets the requirements for detection accuracy in the actual production process,showcasing its effectiveness in improving the precision and speed of metal thin sheet component detection.

关 键 词：机器视觉 金属薄片零件 微小变形 亚像素 缺陷检测 

分 类 号：TG115[金属学及工艺—物理冶金] TP391.41[金属学及工艺—金属学]