基于計算機視覺的模具缺陷自動檢測技術(shù)方法
打開文本圖片集
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
近年來，隨著網(wǎng)絡(luò)技術(shù)的快速發(fā)展，模具質(zhì)量也在不斷提高，以滿足日益增長的技術(shù)需求。（剩余11253字）