作者：刘桂雄, 闫奕樸, 邢星奥

作者单位：华南理工大学机械与汽车工程学院，广东 广州 510640

关键词：无监督学习;表面缺陷;视觉检测

摘要：

视觉表面缺陷检测是工业生产质量控制重要环节，其中无监督学习范式检测方法是重要的发展趋势。该文针对视觉表面缺陷无监督学习检测方法在工业生产、质量控制中的实际应用问题，系统介绍目前国内外的主要物体表面缺陷数据集以及缺陷视觉检测方法主要评价指标，评述图像重建范式、生成模型范式、特征嵌入范式在视觉表面缺陷无监督学习检测中的分类、基本原理及框架、应用性能等方面内容，总结比较各种方法的应用特点以及技术发展趋势，指出归一化流模型、预训练大模型等无监督视觉表面缺陷检测研究值得关注。

Research progress on unsupervised learning detection methods for visual surface defects
LIU Guixiong, YAN Yipu, XING Xing'ao
School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, China
Abstract: Visual surface defect detection is an important part of industrial production and quality control, in which the unsupervised learning paradigm of detection methods is an important development trend. This paper addresses the problem of unsupervised learning detection methods for visual surface defects in industrial production, quality control practical applications, and systematically introduces the current domestic and foreign major object surface defects dataset as well as defects in the visual detection method of the main evaluation index. Reviews the classification, fundamentals and framework, and application performance of the image reconstruction paradigm, generative model paradigm, and feature embedding paradigm for unsupervised learning detection of visual surface defects. This paper summarizes and compares the application characteristics of various methods and the development trends of technology, and points out that research on unsupervised visual surface defect detection such as normalized flow models and pre-trained large models deserves attention.
Keywords: unsupervised learning;surface defect;visual inspection
2024, 50(3):1-12  收稿日期: 2023-11-30;收到修改稿日期: 2024-01-25
基金项目: 广东省重点领域研发计划项目 (2019B010154003)
作者简介: 刘桂雄(1968-),男,广东揭阳市人,教授,研究方向为先进传感与仪器。
参考文献
[1] 赵朗月, 吴一全. 基于机器视觉的表面缺陷检测方法研究进展[J]. 仪器仪表学报, 2022, 43(1): 198-219.
ZHAO L Y, WU Y Q. Research progress of surface defect detection methods based on machine vision[J]. Chinese Journal of Scientific Instrument, 2022, 43(1): 198-219.
[2] 罗东亮, 蔡雨萱, 杨子豪, 等. 工业缺陷检测深度学习方法综述[J]. 中国科学:信息科学, 2022, 52(6): 1002-1039.
LUO D L, CAI Y X, YANG Z H, et al. Survey on industrial defect detection with deep learning[J]. Sci. Sin. Inform, 2022, 52: 1002-1039.
[3] 李少波, 杨静, 王铮, 等. 工业缺陷检测深度学习方法综述[J]. 中国科学: 信息科学, 2022, 52(6): 1002-1039.
LI S, B YANG J, WANG Z, et al. Review of development and application of defect detection technology[J]. Acta Automatica Sinica, 2020, 46(11): 2319-2336.
[4] PENG Y H, YAN Y P, CHEN G Y, et al. Automatic compact camera module solder joint inspection method based on machine vision[J]. Measurement Science and Technology, 2022, 33(10): 105114.
[5] LIU G X, YAN Y P, MENG J. Study on the detection technology for inner-wall outer surface defects of the automotive ABS brake master cylinder based on BM-YOLOv8 [J]. Measurement Science and Technology, 2024, 35(5): 055109.
[6] 徐佳乐, 黄丹平, 廖世鹏, 等. 基于线扫描技术的轴承表面缺陷检测方法研究[J]. 中国测试, 2022, 48(11): 88-94.
XU J L, HUANG D P, LIAO S P, et al. Study on detection method of bearing surface defect based on line scanning technology[J]. China Measurement ＆ Test, 2022, 48(11): 88-94.
[7] 张墩利, 周国栋. 基于改进扩散模型的磁瓦表面缺陷检测方法[J/OL]. 中国测试: 1-7[2024-02-03]. http://kns.cnki.net/kcms/ detail/51.1714.TB.20230413.1936.002.html.
ZHANG D L, ZHOU G D. Surface defects enhancement algorithm of magnetic tile based on improved diffusion model[J/OL]. China Measurement＆Test, 1-7[2024-02-03]. http://kns.cnki.net /kcms/detail/51.1714.TB.20230413.1936.002.html.
[8] CHEN H X, DU Y Z, FU Y Q, et al. DCAM-Net: A rapid detection network for strip steel surface defects based on deformable convolution and attention mechanism[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 1-12.
[9] CHEN W, HE Z L, ZHANG J. Online monitoring of crack dynamic development using attention-based deep networks[J]. Automation in Construction, 2023, 154: 105022.
[10] ELTOUNY K, SAJEDI S, LIANG X. High-resolution vision transformers for pixel-level identification of structural compon-ents and damage[J/OL].  arXiv preprint arXiv. [2020-08-06]. http://doi.org/10.48550/arXiv.2308.03006.
[11] AKHYAR F, LIU Y, HSU C Y, et al. FDD: a deep learning–based steel defect detectors[J]. The International Journal of Advanced Manufacturing Technology, 2023, 126(3-4): 1093-1107.
[12] 何彬媛, 黄坚, 刘桂雄, 等. 面向机箱标准件装配质量局部特征的智能检测技术[J]. 中国测试, 2019, 45(3): 18-23.
HE B Y, HUANG J, LIU G X. Intelligent detection technology for local characteristics of chassis standard component assembly quality[J]. China Measurement ＆ Test, 2019, 45(3): 18-23.
[13] TAO X, MA W Z, LU Z F, et al. Conductive particle detection for chip on glass using convolutional neural network[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 1-10.
[14] YEUNG C C, LAM K M. Efficient fused-attention model for steel surface defect detection[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1-11.
[15] YANG M H, WU P, FENG H. MemSeg: A semi-supervised method for image surface defect detection using differences and commonalities[J]. Engineering Applications of Artificial Intelligence, 2023, 119: 105835.
[16]  LIU J Q, XIE G Y, WANG J B, et al. Deep industrial image anomaly detection: A survey[J]. Machine Intelligence Research, 2024, 21(1): 104-135.
[17] 陶显, 侯伟, 徐德. 基于深度学习的表面缺陷检测方法综述[J]. 自动化学报, 2021, 47(5): 1017-1034.
TAO X, HOU W, XU D. A survey of surface defect detection methods based on deep learning[J]. Acta Automatica Sinica, 2021, 47(5): 1017-1034.
[18] MOHAMMADI B, FATHY M, SABOKROU M. Image/video deep anomaly detection: A survey[J]. arXiv preprint arXiv. [2021-03-02]. http://doi.org/10.48550/arXiv.2103.01739.
[19] TAO X, GONG X Y, ZHANG X, et al. Deep learning for unsuper-vised anomaly localization in industrial images: A survey[J]. IEEE Transactions on Instrumentation and Measure-ment, 2022, 71: 5018021.
[20] BERGMANN P, BATZNER K, FAUSER M, et al. The MVTec anomaly detection dataset: A comprehensive real-world dataset for unsupervised anomaly detection[J]. International Journal of Computer Vision, 2021, 129(4): 1038-1059.
[21]  TSANG C S C, NGAN H Y T, PANG G K H.  Fabric inspection based on the Elo rating method[J]. Pattern Recognit, 2016, 51: 378-394.
[22] HE Y, SONG K C, MENG Q G, et al. An end-to-end steel surface defect detection approach via fusing multiple hierarchical features[J]. IEEE transactions on instrumentation and measure-ment, 2019, 69(4): 1493-1504.
[23] TABERNIK D, ŠELA S, SKVARˇC J, et al. Segmentation-based deep-learning approach for surface-defect detection[J]. Journal of Intelligent Manufacturing, 2020, 31(3): 759-776.
[24] MISHRA P, VERK R, FORNASIER D, et al. VT-ADL: A vision transformer network for image anomaly detec-tion and localization[C]//2021 IEEE 30th International Symposium on Industrial Electronics (ISIE), 2021.
[25] ZOU Y, JEONG J H, PEMULA L, et al. Spot-the-difference self-supervised pre-training for anomaly detection and segmen-tation[C]//European Conference on Computer Vision(ECCV), 2022.
[26] https://scholar.google.com.hk/scholar?hl=zhCN&as_sdt=0%2C5&q=MVTEC-AD&btnG=(2024.01.27).
[27] https://scholar.google.com.hk/scholar?hl=zhCN&as_sdt=0%2C5&q=Fabric+inspection+based+on+the+Elo+rating+method&btnG=(2024.01.27).
[28] https://scholar.google.com.hk/scholar?hl=zhCN&as_sdt=0%2C5&q=An+end-to-end+steel+surface+defect+detection+approach +via+fusing+multiple+hierarchical+features&btnG=(2024.01.27).
[29] https://scholar.google.com.hk/scholar?hl=zhCN&as_sdt=0%2C5&q=Segmentation-based+deep-learning+approach+for+ surface-defect+detection&btnG= (2024.01.27).
[30] https://scholar.google.com.hk/scholar?hl=zhCN&as_sdt=0%2C5&q=VTADL%3A+A+vision+transformer+network+for+image+anomaly+detection+and+localization&btnG= (2024.01.27).
[31] https://scholar.google.com.hk/scholar?hl=zhCN&as_sdt=0%2C5&q=Spot-the-difference+self-supervised+ pretraining+for+ anomaly+detection+and+segmentation&btnG= (2024.01.27).
[32] CUI Y J, LIU Z X, LIAN S G. A survey on unsupervised anomaly detection algorithms for industrial images[J]. IEEE Access, 2023, 11: 55297-55315.
[33] YANG H, CHEN Y F, SONG K Y, et al. Multiscale feature-clustering-based fully convolutional autoencoder for fast accurate visual inspection of texture surface defects[J]. IEEE Transactions on Automation Science and Engineering, 2019, 16(3): 1450-1467.
[34] HE H Y, ZHANG J N, CHEN H X, et al. DiAD: A diffusion-based framework for multi-class anomaly detection[J]. arXiv preprint arXiv.[2023-12-11]. http://doi.org/10.48550/ arXiv. 2312. 06607.
[35] TEBBE J, JAWAD T. D3AD: dynamic denoising diffusion probabilistic model for anomaly detection[J]. arXiv preprint arXiv.[2024-01-09]. http://doi.org/ 10.48550/ arXiv. 2401. 04463.
[36] MICHELUCCI U. An introduction to autoencode-rs[J/OL]. arXiv preprint arXiv. [2022-01-11]. http://doi.org/ 10.48550/arXiv.2201.03898.
[37] YANG L, ZHANG Z L, SONG Y, et al. Diffusion models: A comprehensive survey of methods and applications[J]. ACM Computing Surveys, 2023, 56(4): 1-39.
[38] LIU W Q, LI R Z, ZHENG M, ET al. Towards visually explaining variational autoencoders[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), 2020.
[39] VENKATARAMANAN S, PENG K C, SINGH R V, et al. Attention guided anomaly localization in images[C]//European Conference on Computer Vision(ECCV), 2020.
[40]  YAN X D, ZHANG H D, XU X M, et al. Learning semantic context from normal samples for unsupervised anomaly detection[C]// Proceedings of the AAAI Conference on Artificial Intelligence, 2021.
[41] SONG J W, KONG K B, PARK Y I, et al. AnoSeg: anomaly segmentation network using self-supervised learning[J/OL]. arXiv preprint arXiv. [2021-10-07]. http://doi.org/10.48550 /arXiv.2110.03396.
[42] ALDINUCCI M, DANELUTTO M, KILPATRICK P, et al. Fastflow: High‐level and efficient streaming on multicore[J]. Programming Multi‐core and Many‐core Computing Systems, 2017: 261-280.
[43] GUDOVSKIY D, ISHIZAKA S, KOZUKA K. CFLOW-AD: Real-time unsupervised anomaly detection with localization via conditional normalizing flows[C]// Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision(WACV), 2022.
[44]  ZHOU Y X, XU X, SONG J K, et al. MSFlow: Multiscale flow-based framework for unsupervised anomaly detection[C]//IEEE Transactions on Neural Networks and Learning Systems, 2024.
[45] LIANG Y F, ZHANG J N, ZHAO S W, ET al. Omni-frequency channel-selection representations for unsupervised anomaly detection[J]. IEEE Transactions on Image Processing, 2023, 31: 4327-4340.
[46] RUDOLPH M, BASTIAN W, BODO R. Same same but differnet: Semi-supervised defect detection with normalizing flows[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision(WACV), 2021.
[47] DEFARD T, SETKOV A, LOESCH A, et al. Padim: a patch distribution modeling framework for anomaly detection and localization[C]//International Conference on Pattern Recognition(ICPR), 2021.
[48] ZHANG X, LI S Y, LI X, et al. DeSTSeg: Segmentation guided denoising student-teacher for anomaly detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), 2023.
[49] BATZNER K, HECKLER L, KÖNIG R. EfficientAD: Accurate visual anomaly detection at millisecond-level latencies[C]// Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision(WACV), 2024.
[50] ROTH K, PEMULA L, ZEPEDA J, et al. Towards total recall in industrial anomaly detection[C]//Proc-eedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), 2022.
[51] LIU Z K, ZHOU Y M, XU Y S, et al. Simplenet: A simple network for image anomaly detection and localization[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), 2023.
[52] HYUN J, KIM S, JEON G, et al. Reconpatch: Contrastive patch representation learning for industrial anomaly detection[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision(WACV), 2024.
[53] GU Z P, ZHU B K, ZHU G B, et al. Anomalygpt: Detecting industrial anomalies using large vision-language models[J/OL]. arXiv preprint arXiv. [2023-12-28]. http://doi.org/10.48550/ arXiv.2308.15366.
[54] LI Y Z, WANG H L, YUAN S H, et al. Myriad: large multimodal model by applying vision experts for industrial anomaly detection[J/OL]. arXiv preprint arXiv. [2023-11-01]. http: //doi.org/10.48550/arXiv.2310.19070.
[55] SALEHI M, SADJADI N, BASELIZADEH S, et al. Multiresolution knowledge distillation for anomaly detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), 2021.