摘 要：表面缺陷自动化检测在社会各个行业有广泛应用前景，可以大幅度提升效率。基于卷积神经网络架构的目标检测模型是自动化表面缺陷检测与识别的重要方法。折中检测速度与精确度，选择 tiny-YOLOv3 网络作为表面缺陷检测的模型。将视觉注意力机制引入 tiny-YOLOv3 网络结构并比较不同类别注意力机制在网络不同位置对于模型表现的影响，从而提出一种对于原网络改进的方法。改进的 tiny-YOLOv3 网络结构在表面缺陷数据集上测试结果较原始 tiny-YOLOv3 网络在 mAP 值上提升 2.3%。
关键词：缺陷检测；注意力机制；神经网络
DOI:10.19850/j.cnki.2096-4706.2022.04.025
中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2022）04-0095-05
Research on Surface Defect Detection Based on Improved tiny-YOLOv3 Network
LI Yi 1, WEI Jianguo2, LIU Guanwei 1
(1.Cashway Fintech Co., Ltd., Tianjin 300308, China; 2.College of Intelligence and Computing, Tianjin University, Tianjin 300072, China)
Abstract: Automatic detection of surface defects has a wide application prospect in various industries of society, which can greatly improve the efficiency. The target detection model based on convolutional neural network architecture is an important method for automatic surface defect detection and recognition. To compromise the detection speed and accuracy, choose tiny-YOLOv3 network as the model for surface defect detection. The visual attention mechanism is introduced into the tiny-YOLOv3 network structure and the influence of different types of attention mechanisms in different network positions on the performance of the model are compared, then a method to improve the original network is proposed. The testing results of the improved tiny-YOLOv3 network structure in the surface defect dataset show that the mAP value is 2.3% higher than that of the original tiny-YOLOv3 network.
Keywords: defect detection; attention mechanism; neural network
参考文献：
[1] 黄凤荣，李杨，郭兰申，等 . 基于 Faster R-CNN 的零件表面缺陷检测算法 [J]. 计算机辅助设计与图形学学报，2020，32（6）：883-893.2022.02 99 第4期
[2] 李维刚，叶欣，赵云涛，等 . 基于改进 YOLOv3 算法的带钢表面缺陷检测 [J]. 电子学报，2020，48（7）：1284-1292.
[3] DALAL N，TRIGGS B.Histograms of Oriented Gradients for Human Detection [C]//2005 IEEE Computer Society Conference on Computer Vision & Pattern Recognition.San Diego：IEEE，2005（1）：886-893.
[4] OJALA T，PIETIKAINEN M，MAENPAA T.Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns [J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2002，24（7）：971-987.
[5] LIENHART R，MAYDT J.An Extended Set of Haar-Like Features for Rapid Object Detection [C]//Proceedings.International Conference on Image Processing.Rochester：IEEE，2002（1）：I-I.
[6] SUYKENS J，VANDEWALLE J.Least Squares Support Vector Machine Classifiers [J].Neural Processing Letters，1999，9（3）：293-300.
[7] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation [C]//2014 IEEE Conference on Computer Vision and Pattern Recognition.Columbus：IEEE，2014，580-587.
[8] GIRSHICK R.Fast R-CNN [C]//2015 IEEE International Conference on Computer Vision（ICCV）.Santiago：IEEE，2015：1440-1448.
[9] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：Towards Real-Time Object Detection with Region Proposal Networks [J]. IEEE Transactions on Pattern Analysis & Machine Intelligence，2017，39（6）：1137-1149.
[10] REDMON J，DIVVALA S，GIRSHICK R，et al.You Only Look Once：Unified，Real-Time Object Detection [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）. Las Vegas：IEEE，2016，779-788.
[11] REDMON J，FARHADI A.YOLO9000：Better，Faster，Stronger [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）.Honolulu：IEEE，2017：6517-6525.
[12] REDMON J，FARHADI A.YOLOv3：An Incremental Improvement [J/OL].arXiv：1804.02767 [cs.CV].[2021-12-03].https://arxiv.org/abs/1804.02767.
[13] LIU W，ANGUELOV D，ERHAN D，et al.SSD：Single Shot MultiBox Detector [C]//Computer Vision – ECCV 2016.Cham：Springer，2016：21–37.
[14] JIE H，LI S，GANG S.Squeeze-and-Excitation Networks [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake City：IEEE，2018：7132-7141.
[15] WOO S，PARK J，LEE J Y，et al.CBAM：Convolutional Block Attention Module [J/OL].arXiv：1807.06521 [cs.CV].[2021-11-03].https://arxiv.org/abs/1807.06521.
[16] YU J，JIANG Y，WANG Z，et al.UnitBox：An Advanced Object Detection Network [C]//MM’16：Proceedings of the 24th ACM international conference on Multimedia.Amsterdam：Association for Computing Machinery，2016：516–520.
[17] REZATOFIGHI H，TSOI N，GWAK J Y，e t al.Generalized Intersection Over Union：A Metric and a Loss for Bounding Box Regression [C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）.Long Beach：IEEE，2019：658-666.
[18] ZHENG Z，WANG P，LIU W，et al.Distance-IoU Loss:Faster and Better Learning for Bounding Box Regression [J/OL].arXiv：1911.08287 [cs.CV].[2021-11-13].https://arxiv.org/abs/1911.08287.
[19] KINGMA D，BA J.ADAM：A Method for Stochastic Optimization [J/OL].arXiv：1412.6980 [cs.LG].[2021-11-23].https://arxiv.org/abs/1412.6980.
作者简介：李屹（1984—），男，汉族，山东滨州人，工程师，博士研究生，研究方向：图像处理、计算机视觉。