基于四焦距相位相干机器视觉的晶体微缺陷检测  

Crystal Micro Defect Detection Based on Four Focal Length Phase-Coherent Machine Vision

作　　者：朱涛 黄玉玲 陶昕辰 张泽宇 朱厚森 陈王宇 羊箭锋 杨勇 吴迪 ZHU Tao;HUANG Yuling;TAO Xinchen;ZHANG Zeyu;ZHU Housen;CHEN Wangyu;YANG Jianfeng;YANG Yong;WU Di(School of Optoelectronic Science and Engineering,Soochow University,Suzhou Jiangsu 215006,China)

机构地区：[1]苏州大学光电科学与工程学院,江苏苏州215006

出　　处：《电子器件》2024年第3期743-748,共6页Chinese Journal of Electron Devices

基　　金：苏州大学“大学生创新创业训练计划”项目(202110285074K)。

摘　　要：针对目前在晶体内部微缺陷研究上的检测精度较低、检测尺度不足的问题,提出了一种基于四焦距相位相干机器视觉的晶体内部微缺陷检测法,并建立了相应的检测系统。本方法通过四焦距相位相干光路,将晶体内部缺陷的相位信息转化为灰度信息,形成晶体内部缺陷特征图像。利用机器视觉算法,提取缺陷特征图像中的相位信息丢失点,并通过滤波算法对这些点进行恢复。然后通过特征提取、灰度分析算法,对缺陷特征图像进行处理分析,最后得出晶体内部缺陷的厚度,完成对晶体内部微缺陷的检测。实验结果表明,利用该方法对有机玻璃表面中心镀60 nm的二氧化硅膜的模拟相位缺陷进行检测,相对误差为1.83%,检测分辨率达到纳米量级。相对于动态泰曼干涉仪检测的2.4%的相对误差以及激光聚焦线扫描法的40μm检测分辨率,有着较大的提升,也证明所提出的方法能够有效运用于晶体内部的微缺陷的检测。Focusing on the problem of low detection accuracy in the minor internal defect detection of crystals,a method based on four focal length phase-coherent machine vision is proposed.The phase information of crystal internal defects is turned into gray level through four focal length phase-coherent optical path.Then a feature image of crystal internal defects is formed.The missing points in the feature image can be extracted by using machine vision algorithm.After feature extraction and gray analysis of the feature image,the thickness of crystal internal defects is obtained.The simulated defect of 60 nm silica film coated on the center of plexiglass surface is tested,the results show that the relative error of the method is 1.83%and the detection resolution reaches the order of nanometers.Compared with the relative error of 2.4%of the dynamic Teman interferometer and the 40μm detection resolution of laser focusing line scanning method,it has good improvements.The results also prove that the method proposed can be effectively applied to the detection of minor defects in the crystal.

关 键 词：图像处理 相位相干 机器视觉 晶体 缺陷检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]