模具技术Issue(2)：85-93,9.
基于计算机视觉的模具缺陷自动检测技术方法
Automatic mold defect detection technology based on competer vision
摘要
Abstract
In order to improve mold quality and enhance the efficiency and accuracy of mold defect detection,this paper proposes a computer vision-based automatic mold defect detection technology.First,the model space feature data is collected,and the initial particle positions and velocities in the population are set.Then,an adaptive function is used to analyze and find the local and global optimal solutions,optimizing the particle velocities and positions.Next,the Sigmoid function is applied to adjust the particle velocities and positions.Finally,the stability of the particle's fitness is evaluated,and the mold defect detection results are visualized and output by a computer.Experimental results show that the proposed technology can effectively identify defect types,contents,and sizes,and the detection time is reduced by at least 55％compared to traditional manual detection methods.Moreover,the technology has a false detection rate of 1.18％,a true detection rate of 99.32％,and a missed detection rate of 8.76％,demonstrating the best performance among all detection technologies.关键词
模具缺陷检测/计算机视觉/适应度函数/Sigmoid函数/自动检测技术Key words
mold defect detection/computer vision/fitness function/sigmoid function/automatic detection technology分类
信息技术与安全科学引用本文复制引用
吕斌,王艳..基于计算机视觉的模具缺陷自动检测技术方法[J].模具技术,2025,(2):85-93,9.基金项目
陕西省教育厅一般专项科学研究计划项目"管道缺陷检测机器人高度集成与智能化研究"(编号:24JK0383). （编号:24JK0383）
