基于计算机视觉的模具缺陷自动检测技术方法

打开文本图片集

Automatic mold defect detection technology based on competer vision

LUBin，WANG Yan

（Shaanxi Energy Institute， Xianyang 712Ooo， Shaanxi，China）

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 hasa falsedetection rate of 1.18% ，atrue detection rate of 99.32% ，anda missed detection rate of 8.76% ，demonstrating the best performance among all detection technologies.

Key words： mold defect detection; computer vision； fitness function; sigmoid function; automatic detection technology

0 引言

近年来，随着网络技术的快速发展，模具质量也在不断提高，以满足日益增长的技术需求。（剩余11253字）