[1]赖肖强,胡 泓.基于机器视觉算法和改进 MobileNetV2 的晶圆缺陷检测研究[J].机械与电子,2025,(01):34-39.
　LAI Xiaoqiang,HU Hong.Research on Wafer Defect Detection Based on Machine Vision Algorithms and Improved MobileNetV2[J].Machinery & Electronics,2025,(01):34-39.
点击复制

基于机器视觉算法和改进 MobileNetV2 的晶圆缺陷检测研究
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
期数:
2025年01期
页码:
34-39
栏目:
自动控制与检测
出版日期:
2025-01-30

文章信息/Info

Title:
Research on Wafer Defect Detection Based on Machine Vision Algorithms and Improved MobileNetV2
文章编号:
1001-2257 ( 2025 ) 01-0034-06
作者:
赖肖强; 胡 泓
哈尔滨工业大学(深圳)机电工程与自动化学院深圳市先进制造重点实验室,广东 深圳 518063
Author(s):
LAI Xiaoqiang ;  HU Hong
( Shenzhen Key Laboratory of Advanced Manufacturing , School of Mechanical and Electrical Engineering and Automation , Harbin Institute of Technology ( Shenzhen ), Shenzhen 518063 , China )
关键词:
晶圆; 缺陷检测; 机器视觉;  AOI ; 图像处理
Keywords:
wafer ;  defect detection ;  machine vision ;  automatic optical inspection ;  image processing
分类号:
TP391.4
文献标志码:
A
摘要:
基于机器视觉以及 AOI 自动光学检测平台,设计了一套针对晶圆表面缺陷的自动化检测系统。首先设计了晶圆缺陷检测的总体方案,分别对 AOI 系统的机械结构以及运动控制模块制定合理方案。然后基于机器视觉算法设计晶圆缺陷检测的算法,用于检测晶圆表面的光敏面脏污、光敏面划痕和电极环断开缺陷。同时基于改进后的 MobileNetV2 模型实现对晶圆表面电极脖子区域的缺陷检测。这一套方法可以实现晶粒的缺陷检测,脏污、划痕、电极环断开等缺陷的检测准确率分别达到 94.06% 、 97.37% 、 92.36% ,平均单颗晶粒检测耗时分别为 86 ms 、 13 ms 、 93 ms ,使用基于改进 MobileNetV2 模型的电极脖子缺陷检测准确率达到 99.3% 。以上实验结果表明,所提方法准确率高、耗时短,可以满足晶圆缺陷检测的需求。
Abstract:
This paper introduces an automated defect detection system for wafer surfaces , combining machine vision and AOI technology.The overall approach includes designing mechanical and motion control modules for the AOI system , implementing machine vision algorithms to detect various wafer surface defects , and employing an enhanced MobileNetV2 model for electrode neck defect detection.The system achieves high accuracy rates of 94.06% , 97.37% , and 92.36% for detecting contamination , scratches , and electrode ring disconnection , respectively.The average detection times per wafer grain are 86 ms , 13 ms , and 93 ms for these defects.Additionally , the electrode neck defect detection accuracy using the improved MobileNetV2 model reaches 99.3%.These results demonstrate the proposed method ’ s effectiveness in achie-ving accurate and efficient wafer defect detection.

参考文献/References:

[ 1 ] BRYANT R , HICKSON J , HAWTHORNE J.Novelfull wafer inspection technology for non-visual residue defects [ J ] .Solid state phenomena , 2008 , 134 : 289-292.

[ 2 ] 陈世炜 . 基于明暗场成像的多扫描方式图案化晶圆检测技术研究[ D ] . 杭州:浙江大学,2021.
[ 3 ] 张俊 . 基于机器视觉的线缆表观缺陷在线检测系统的研究[ D ] . 成都:电子科技大学,2017.
[ 4 ] BUSTACARA-MEDINA C , FLOREZ-VALENCIA L , DIAZ L C.Improved canny edge detector using principal curvatures [ J ] .Journal ofelectrical and electronic engineering , 2020 , 8 ( 4 ): 109-116.
[ 5 ] WANG R , CHEN N.Wafer map defect pattern recognition using rotation invariant features [ J ] .IEEE Transactions onsemiconductor manufacturing , 2019 , 32( 4 ): 596-604.
[ 6 ] WRIGHT C , YANG S J.Deep learning for automated focus quality detection in wafer inspection [ C ] ∥Pro- ceedings of the Automated Visual Inspection and Machine Vision IV , 2021 : 2592425.
[ 7 ] 顾寄南,唐良颖,许悦,等 . 机器视觉在网片缺陷检测与分类中的应用[ J ] . 机械设计与制造, 2019(增刊 1 ):47-49 , 53.
[8 ] 葛文超 . 基于半监督流形学习的晶圆表面缺陷检测方法研究[ D ] . 合肥:合肥工业大学,2021.
[ 9 ] 韩茜茜,耿世勇,李恒毅 . 基于机器视觉的缺陷检测应用综述[ J ] . 电工技术, 2019 ( 14 ): 117-118 , 132.
[ 10 ] KIM H T , KIM S T , CHO Y J.Quick and efficient light control for conventional automatic optical inspection ( AOI ) systems [ J ] .International journal of precision engineering and manufacturing , 2015 , 16 :247-254.?
[ 11 ] CHEN L C , CHANG Y W.Innovative simultaneous confocal full field 3D surface profilometry for in situ automatic optical inspection ( AOI )[ J ] .Measurement science and technology , 2010 , 21 ( 6 ): 065301.
[ 12 ] 张翠芳,杨国为,岳明明 .Zhang 并行细化算法的改进[ J ] . 信息技术与信息化, 2016 ( 6 ): 69-71.
[ 13 ] 牟少敏,杜海洋,苏平 . 一种改进的快速并行细化算法[ J ] . 微电子学与计算机, 2013 , 30 ( 1 ): 53-55 , 60.
[ 14 ] 李王辉,肖文珂 . 基于 ARM 平台的手写体数字快速识别方法[ J ] . 信息与电脑(理论版), 2022 , 34 ( 7 ): 201-203.
[15 ] 郭毅强 . 晶圆表面缺陷视觉检测研究[ D ] . 哈尔滨:哈尔滨工业大学,2020.
[ 16 ] HU J , SHEN L , SUN G.Squeeze and excitation networks [ C ] ∥2018 IEEE / CVFConference on Computer Vision and Pattern Recognition , 2018 : 7132-7141.

相似文献/References:

[1]陆 晓,高 超,蒋承伶,等.基于双注意力机制和级联检测框架的金具螺栓缺陷检测研究[J].机械与电子,2024,42(01):63.
　LU Xiao,GAO Chao,JIANG Chengling,et al.Research on Bolt Defect Detection Based on Dual-attention Mechanism and Cascade Detection Framework[J].Machinery & Electronics,2024,42(01):63.
[2]黄心玥,王明泉,耿宇杰,等.基于 YOLOv8 的铝合金铸造轮毂缺陷检测算法研究与分析[J].机械与电子,2025,(02):26.
　HUANG Xinyue,WANG Mingquan,GENG Yujie,et al.Research and Analysis of Defect Detection Algorithm for Aluminum Alloy Casting Wheels Based on YOLOv8[J].Machinery & Electronics,2025,(01):26.
[3]曹振锋,王明泉,路宇鹏,等.基于 YOLOv8 的轻量化轮毂缺陷检测算法研究[J].机械与电子,2025,(06):31.
　CAO Zhenfeng,WANG Mingquan,LU Yupeng,et al.Research on the Lightweight Wheel Hub Defect Detection Algorithm Based on YOLOv8[J].Machinery & Electronics,2025,(01):31.
[4]刘 骞,陈茂林.基于深度学习的少样本光伏边框划痕检测[J].机械与电子,2025,(08):47.
　LIU Qian,CHEN Maolin.Few-shot Scratch Detection on Photovoltaic Frame Based on Deep Learning[J].Machinery & Electronics,2025,(01):47.
[5]王立杰,高嘉伟,徐相龙,等. 基于SDA-YOLOv8的PCBA 缺陷检测算法研究[J].机械与电子,2026,44(01):45.
　WANG Lijie,GAO Jiawei,XU Xianglong,et al. Research on PCBA Defect Detection Algorithm Based on SDA -YOLOv8[J].Machinery & Electronics,2026,44(01):45.

备注/Memo

备注/Memo:
收稿日期: 2023-12-29
基金项目:深圳市科技计划资助项目( JSGG20201201100410029 )
作者简介:赖肖强 ( 1999- ),男,福建漳州人,硕士研究生,研究方向为机器视觉;胡 泓( 1965- ),男,天津人,博士,教授,研究方向为微机电系统( MEMS )、微流控技术与精密仪器、线性和非线性控制系统理论、机电一体化系统与智能系统,通信作者。
更新日期/Last Update: 2025-03-06