基于灰狼优化算法的改进Canny算子的芯片标识图像边缘检测  

Chip Marking Image Edge Detection Based on Improved Canny Operator of Grey Wolf Optimization Algorithm

作　　者：刘勍 郝静 侯喆 赵利民 赵玉祥 张进兵 LIU Qing;HAO Jing;HOU Zhe;ZHAO Limin;ZHAO Yuxiang;ZHANG Jinbing(School of Electronic Information&Electrical Engineering,Tianshui Normal University,Tianshui 741001,China;Engineering Research Center of Integrated Circuit Packaging and Testing,Ministry of Education,Tianshui 741001,China;Tianshui Huatian Technology Group Co.,Ltd.,Tianshui 741000,China)

机构地区：[1]天水师范学院电子信息与电气工程学院,甘肃天水741001 [2]集成电路封装测试教育部工程研究中心,甘肃天水741001 [3]天水华天电子集团股份有限公司,甘肃天水741000

出　　处：《贵州大学学报（自然科学版）》2024年第5期41-48,共8页Journal of Guizhou University:Natural Sciences

基　　金：国家自然科学基金资助项目(61461046);甘肃省自然科学基金资助项目(20JR10RA802,20JR5RA494);甘肃省科技重大专项计划资助项目(22ZD6GE016,23ZDGE001);甘肃省教育厅教育揭榜挂帅项目(2021jyjbgs-06);天水师范学院科研项目(PTJ2022-01,PTJ2022-04);天水市秦州区科技计划资助项目(2023-SHFZG-6476);天水师范学院研究生创新引导项目(2023CXZX-802,TYCX2236);甘肃省2023年度重点人才项目(2023RCXM29)。

摘　　要：为有效进行芯片标识的提取,提出一种基于灰狼优化算法(gray wolf optimization,GWO)的改进动态双阈值的Canny算子来进行芯片标识图像边缘提取。首先,从芯片标识生产环境复杂、图像干扰信息多的角度出发,对Canny算子的双阈值进行改进;其次,使用灰狼优化算法确定其高阈值选取;最后,将本文算法与传统Log、Prewitt、Roberts、Canny、Sobel算子进行实验比较,利用召回率和精确率等方法作了客观评估。实验结果表明,本文所提算法优于传统的边缘提取算法,提取准确度高,为后续识别打下了坚实基础。To effectively extract chip marking,an improved dynamic dual threshold Canny operator based on gray wolf optimization(GWO)algorithm is proposed for edge extraction of chip marking images.Firstly,from the perspective of the complex image interference information in the chip marking production environment,the dual threshold of the Canny operator is improved.Secondly,the grey wolf algorithm is used to determine its high threshold selection;Finally,the algorithm proposed in this paper is experimentally compared with traditional Log operators,traditional Prewitt operators,traditional Roberts operators,traditional Canny operators,and traditional Sobel operators,and objective evaluations are conducted using methods such as recall and accuracy.The experimental results show that the algorithm proposed in this study is superior to traditional edge extraction algorithms,with high extraction accuracy,and lays a solid foundation for subsequent recognition tasks.

关 键 词：芯片标识图像 边缘检测 改进Canny算子 GWO 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]