基于网格的多目标模板匹配晶粒位置检测方法  

Grid-based Multi-targets Template Matching Die Position Detection Method

作　　者：周书辰 陈晓荣[1] 王子旋 ZHOU Shuchen;CHEN Xiaorong;WANG Zixuan(School of Optoelectronic Information and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《计量学报》2024年第5期639-645,共7页Acta Metrologica Sinica

基　　金：国家自然科学基金(52175513)。

摘　　要：针对传统方法中晶粒位置检测的耗时长、精度低等局限性,提出一种基于网格的多目标模板匹配晶粒位置检测方法。通过改进传统的模板匹配方法,结合非极大值抑制算法,将芯片晶粒的检测速度和精度提高。实验结果表明:该算法在单一同种晶粒的算法识别率能够达到97%以上,单张图像耗时<200 ms,能够克服明暗不同造成的检测困难,达到技术指标要求。To address the limitations of traditional methods in terms of long detection time and low accuracy,a multi-objective template matching algorithm detecting position of die based on grid is proposed.By improving the traditional template matching method and combining with non-maximum suppression algorithm,the detection speed and accuracy of chip dice are improved.Experimental results show that the proposed algorithm has a recognition rate of over 97%for a single same die,and the single image processing time does not exceed 200 ms.It can overcome the detection difficulties caused by different brightness and meet the technical specifications.

关 键 词：光电检测 晶粒位置检测 机器视觉 非极大值抑制 网格 多目标模板匹配 

分 类 号：TB96[机械工程—光学工程] TB973[一般工业技术—计量学]