基于X光图像的芯片在线自适应计数算法研究  

Online Adaptive Counting Algorithm for Chips Based on X Optical Images

作　　者：雷臻宇 贺香华 骆佩文 周聪 谢德芳 LEI Zhenyu;HE Xianghua;LUO Peiwen;ZHOU Cong;XIE Defang(Mechanical and Electrical Engineering College,Guangzhou University,Guangzhou 510006,China)

机构地区：[1]广州大学机械与电气工程学院,广州510006

出　　处：《机电工程技术》2020年第2期80-84,共5页Mechanical & Electrical Engineering Technology

摘　　要：为提高生产线上芯片计数的自动化程度,提出一种基于X光图像的芯片在线自适应计数算法。对X光拍摄的半导体芯片图像进行预处理,利用自动阈值提炼背景区域;对提炼后的背景区域进行自动阈值提取出芯片区域,通过遍历芯片区域的面积估算单个芯片的面积;设置相应的倍数关系来区分粘连区域和噪声区域,将所有连通区域的个数求和即可得到图像中所有芯片的个数。测试结果表明,自动计数得出的数量偏差率平均为0.03%,耗时平均2.9s 1张,能满足生产线上自动计数的精度高、耗时短等要求。In order to improve the automation of chip counting in production line,an on-line adaptive chip counting algorithm based on X-ray image is proposed.Preprocess the semiconductor chip image captured by X-ray,extract the background area by automatic threshold,extract the chip area by automatic threshold,estimate the area of a single chip by traversing the area of the chip area,set the corresponding multiple relationship to distinguish the adhesion area and noise area,and sum the number of all the connected areas to get the image number of chips.The test results show that the average deviation rate of quantity obtained by automatic counting is 0.03%,and the average time-consuming is 2.9 s.It can meet the requirements of high precision and short time-consuming of automatic counting in production line.

关 键 词：芯片 自动计数 图像分割 面积特征 参考面积 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]