摘要
当前检测多通道数字集成芯片缺陷时,存在清晰度低、疲劳极限精度低以及检测精度低的问题,提出多通道数字集成芯片缺陷的线激光锁相热成像检测方法。利用可视化算法,通过激光锁相热成像方法采集多通道数字集成芯片图像,采用Holder指数运算锁相幅度图像中所有断连的图像,利用激光扫描再次进行图像成像处理,通过主成分分析法提取不连续图像的特征,实现芯片缺陷的检测。结果表明,本方法在第10秒完成清晰成像,疲劳极限精度较高,对芯片进行缺陷检测的精度最高,证明了本方法在实际应用中具有较好的实用性。
Currently,when detecting defects in multi-channel digital integrated chips,there are problems of low definition,low fatigue limit accuracy,and low detection accuracy.A line laser phase-locked thermal imaging detection method for defects in multi-channel digital integrated chips is proposed.Use visualization algorithms to collect multi-channel digital integrated chip images by laser phase-locked thermal imaging method,use Holder index to calculate all disconnected images in the phase-locked amplitude image,use laser scanning to perform image imaging again,and extract the images by principal component analysis.The characteristics of continuous images enable the detection of chip defects.The results show that the method in this paper completes clear imaging in the 10 th second,the fatigue limit accuracy is high,and the accuracy of chip defect detection is the highest,which proves that the method in this paper has good practicability in practical applications.
作者
张稼
陆兴华
祝振宇
ZHANG Jia;LU Xinghua;ZHU Zhenyu(Huali College Guangdong University of Technology,Guangzhou 511300,China)
出处
《激光杂志》
CAS
北大核心
2022年第8期76-80,共5页
Laser Journal
基金
广东省教育厅特色创新项目(No.2018WTSCX204)
广东省教育科学“十三五”规划项目(No.2020GXJK479)。
关键词
芯片缺陷
激光锁相
图像叠加
特征提取
chip defect
laser phase locked
image superposition
feature extraction
