基于机器视觉的芯片字符区域分割和定位算法  

Chip Character Area Segmentation and Localization Algorithm Based on Machine Vision

作　　者：陈甦欣[1] 赵安宁 罗乐文 CHEN Suxin;ZHAO Anning;LUO Lewen(School of Mechanical Engineering,Hefei University of Technology,Hefei 230009,China)

机构地区：[1]合肥工业大学机械工程学院,合肥230009

出　　处：《组合机床与自动化加工技术》2024年第4期10-13,18,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：安徽省科技重大专项资金项目(2021d05050002)。

摘　　要：芯片表面的字符对分选具有重要意义,对字符的定位是分选工作的关键步骤。为了提高分选工作定位效率和定位精度,提出了一种基于改进的区域生长算法和凸包检测算法分割和定位芯片表面字符区域方法。首先,对采集图像进行预处理操作,利用改进的Canny算法获取无干扰图像边缘,将图像边缘作为区域生长法的种子点并以图像自适应阈值作为生长准则分割图像,使用最大内接矩形算法粗定位字符区域;其次,采用Harris角点检测算法获取字符角点分布位置;最后,筛选角点并提取关键点,利用一种改进的凸包检测算法定位字符区域。经过实验验证,所设计的算法能够完整的分割和定位芯片表面字符区域,定位精度和效率较原有方法提升5.3%和15.4%,满足实际工业生产的要求。The characters on the chip surface are important for sorting,and the positioning of the characters is a key step in the sorting work.In order to improve the localization efficiency and localization accuracy of the sorting work,a method is proposed to segment and localize the character area on the chip surface based on the improved region growing algorithm and convex packet detection algorithm.Firstly,we perform pre-processing operations on the acquired images;use the improved Canny algorithm to obtain the interference-free image edges,use the image edges as the seed points of the region growth method and segment the ima-ges with the image adaptive threshold as the growth criterion;use the maximum inline rectangle algorithm to coarsely locate the character regions;then use the Harris corner detection algorithm to obtain the charac-ter corner distribution locations;finally,filter the corner points and extract the key points Finally,a modified convex packet detection algorithm is used to locate the character region.After experimental verification,the designed algorithm can completely segment and locate the character area on the chip surface,and the posi-tioning accuracy and efficiency are improved by 5.3%and 15.4%compared with the original method,meeting the requirements of actual industrial production.

关 键 词：机器视觉 区域生长法 角点检测 凸包检测 字符定位 

分 类 号：TH161[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]