[1]杨奡飞,续欣莹,谢刚,等.印刷电路板缺陷持续检测与定位方法研究[J].智能系统学报,2025,20(1):219-229.[doi:10.11992/tis.202310024]
YANG Aofei,XU Xinying,XIE Gang,et al.Research on continual detection and localization method for printed circuit board defect[J].CAAI Transactions on Intelligent Systems,2025,20(1):219-229.[doi:10.11992/tis.202310024]
点击复制
YANG Aofei,XU Xinying,XIE Gang,et al.Research on continual detection and localization method for printed circuit board defect[J].CAAI Transactions on Intelligent Systems,2025,20(1):219-229.[doi:10.11992/tis.202310024]
印刷电路板缺陷持续检测与定位方法研究
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
20
期数:
2025年第1期
页码:
219-229
栏目:
吴文俊人工智能科学技术奖论坛
出版日期:
2025-01-05
- Title:
- Research on continual detection and localization method for printed circuit board defect
- Keywords:
- defect detection; defect localization; continual learning; deep learning; unsupervised learning; reverse distillation; one-class classification embedding; pooling distillation; printed circuit board
- 分类号:
- TP391
- 摘要:
- 针对目前缺陷检测与定位方法只能对特定类型的缺陷进行检测，而不能连续地学习检测不同类型缺陷的问题，提出了一种基于反向蒸馏模型的缺陷检测与定位方法。该方法以反向蒸馏模型为基础模型，对模型中间层输出的特征图以及一分类嵌入表示进行池化蒸馏，使得模型能够在连续的任务序列上不断地学习新的检测任务，从而达到持续学习的能力。在4个连续的印刷电路板（printed circuit board, PCB）缺陷检测与定位任务上进行实验，实验结果表明该方法的性能优于对比方法，能够满足工业生产场景的应用需求，在抑制对旧任务样本的检测能力的遗忘的同时，能够保持学习检测新任务的能力。
- Abstract:
- Existing defect detection and localization methods can only detect fixed types of defects and cannot meet the continual defect detection requirements in real application scenarios. To address this issue, this paper proposes a defect detection and localization method based on the reverse distillation model. This method uses the reverse distillation model as the basis model and performs pooling distillation on the feature maps from the middle layers of the model and the one-class classification embedding representation. So that the model can continually train new detection tasks without forgetting previous tasks. Experimental results on four printed circuit board defect detection and localization tasks show that this method can meet the requirements of industrial applications, and it outperforms other methods. It maintains the ability to learn and detect new tasks while suppressing the trend of forgetting the ability to detect samples of previous tasks.
- 参考文献/References:
-
[1] DING Runwei, DAI Linhui, LI Guangpeng, et al. TDD-net: a tiny defect detection network for printed circuit boards[J]. CAAI transactions on intelligence technology, 2019, 4(2): 110-116.
[2] ADIBHATLA V A, CHIH H C, HSU C C, et al. Defect detection in printed circuit boards using you-only-look-once convolutional neural networks[J]. Electronics, 2020, 9(9): 1547.
[3] LIN Qiangqiang, ZHOU Jinzhu, MA Qiurui, et al. EMRA-net: a pixel-wise network fusing local and global features for tiny and low-contrast surface defect detection[J]. IEEE transactions on instrumentation measurement, 2022, 71: 3151926.
[4] 钱万明, 朱红萍, 朱泓知, 等. 基于自适应加权特征融合的PCB裸板缺陷检测研究[J]. 电子测量与仪器学报, 2022, 36(10): 92-99.
QIAN Wanming, ZHU Hongping, ZHU Hongzhi, et al. Research on defect detection of PCB bare board based on adaptive weighted feature fusion[J]. Journal of electronic measurement and instrumentation, 2022, 36(10): 92-99
[5] 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 Xian, HOU Wei, XU De. A survey of surface defect detection methods based on deep learning[J]. Acta automatica sinica, 2021, 47(5): 1017-1034.
[6] RUFF L, VANDERMEULEN R, GOERNITZ N, et al. Deep one-class classification[C]//Proceedingds of the 35th International conference on machine learning. Stockholm: PMLR, 2018: 4393-4402.
[7] LIZNERSKI P, RUFF L, VANDERMEULEN R A, et al. Explainable deep one-class classification[EB/OL]. (2020-07-03)[2023-10-19]. http://arxiv.org/abs/2007.01760v3.
[8] BERGMAN L, HOSHEN Y. Classification-based anomaly detection for general data[EB/OL]. (2020-05-05) [2023-10-19]. http://arxiv.org/abs/2005.02359v1.
[9] COHEN N, HOSHEN Y. Sub-image anomaly detection with deep pyramid correspondences[EB/OL]. (2020-05-05) [2023-10-19]. http://arxiv.org/abs/2005.02357v3.
[10] 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. Cham: Springer, 2021: 475-489.
[11] ROTH K, PEMULA L, ZEPEDA J, et al. Towards total recall in industrial anomaly detection[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans: IEEE, 2022: 14298-14308.
[12] CRESWELL A, WHITE T, DUMOULIN V, et al. Generative adversarial networks: an overview[J]. IEEE signal processing magazine, 2018, 35(1): 53-65.
[13] 陈大鹏, 姚剑敏, 严群, 等. 基于自编码器结构改进的无监督图像异常检测[J]. 信息技术与信息化, 2023(8): 4-7.
CHEN Dapeng, YAO Jianmin, YAN Qun, et al. Unsupervised image anomaly detection based on improved autoencoder structure[J]. Information technology and informatization, 2023(8): 4-7.
[14] WANG Guodong, HAN Shumin, DING Errui, et al. Student-teacher feature pyramid matching for anomaly detection[EB/OL]. (2021-03-07)[2023-10-19]. http://arxiv.org/abs/2103.04257v3.
[15] BERGMANN P, FAUSER M, SATTLEGGER D, et al. Uninformed students: student-teacher anomaly detection with discriminative latent embeddings[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle: IEEE, 2020: 4182-4191.
[16] 秦荣荣, 高晓蓉, 罗林, 等. 基于注意力反向知识蒸馏的车轮踏面异常检测[J]. 激光与光电子学进展, 2023, 60(24): 230-239.
QIN Rongrong, GAO Xiaorong, LUO Lin, et al. Wheel tread anomaly detection based on attentional reverse knowledge distillation[J]. Laser & optoelectronics progress, 2023, 60(24): 230-239.
[17] KIRKPATRICK J, PASCANU R, RABINOWITZ N, et al. Overcoming catastrophic forgetting in neural networks[J]. Proceedings of the national academy of science, 2017, 114(13): 3521-3526.
[18] LI Zhizhong, HOIEM D. Learning without forgetting[J]. IEEE transactions on pattern analysis and machine intelligence, 2018, 40(12): 2935-2947.
[19] 张静宇, 续欣莹, 谢刚, 等. 基于弹性权重巩固与知识蒸馏的垃圾持续分类[J]. 智能系统学报, 2023, 18(4): 878-885.
ZHANG Jingyu, XU Xinying, XIE Gang, et al. Continuous classification of garbage based on the elastic weight consolidation and knowledge distillation[J]. CAAI transactions on intelligent systems, 2023, 18(4): 878-885.
[20] LIU Hao, LIU Huaping. Continual learning with recursive gradient optimization[EB/OL]. (2022-01-29)[2023-10-19]. http://arxiv.org/abs/2201.12522v1.
[21] REBUFFI S A, KOLESNIKOV A, SPERL G, et al. iCaRL: incremental classifier and representation learning[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu: IEEE, 2017: 5533-5542.
[22] BUZZEGA P, BOSCHINI M, PORRELLO A, et al. Dark experience for general continual learning: a strong, simple baseline[EB/OL]. (2020-04-15)[2023-10-19]. http://arxiv.org/abs/2004.07211v2.
[23] HU Zhiyuan, LI Yunsheng, LYU Jiancheng, et al. Dense network expansion for class incremental learning[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver: IEEE, 2023: 11858-11867.
[24] MALLYA A, LAZEBNIK S. PackNet: adding multiple tasks to a single network by iterative pruning[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018: 7765-7773.
[25] FERNANDO C, BANARSE D, BLUNDELL C, et al. PathNet: evolution channels gradient descent in super neural networks[EB/OL]. (2017-01-30)[2023-10-19]. http://arxiv.org/abs/1701.08734v1.
[26] 陈焕文. 网络分割与自适应突触可塑性融合的连续学习[J]. 华中科技大学学报(自然科学版), 2024, 52(3): 156-160.
CHEN Huanwen. Continual learning using split network and adaptive synaptic plasticity[J]. Journal of huazhong university of science and technology(natural science edition), 2024, 52(3): 156-160.
[27] PEZZE D D, ANELLO E, MASIERO C, et al. Continual learning approaches for anomaly detection[EB/OL]. (2022-11-21)[2023-10-19]. http://arxiv.org/abs/2212.11192v1.
[28] DENG Hanqiu, LI Xingyu. Anomaly detection via reverse distillation from one-class embedding[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans: IEEE, 2022: 9727-9736.
[29] DOUILLARD A, CORD M, OLLION C, et al. PODNet: pooled outputs distillation for small-tasks incremental learning[C]//European Conference on Computer Vision. Cham: Springer, 2020: 86-102.
[30] KINGMA D P, BA J, HAMMAD M M. Adam: a method for stochastic optimization[EB/OL]. (2014-12-22)[2023-10-19]. http://arxiv.org/abs/1412.6980v9.
[31] TIEN T D, NGUYEN A T, TRAN N H, et al. Revisiting reverse distillation for anomaly detection[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver: IEEE, 2023: 24511-24520.
[32] ZAVRTANIK V, KRISTAN M, SKO?AJ D. DRAEM-A discriminatively trained reconstruction embedding for surface anomaly detection[C]//2021 IEEE/CVF International Conference on Computer Vision. Montreal: IEEE, 2021: 8310-8319.
- 相似文献/References:
-
[1]吴家伟,严京旗,方志宏,等.基于图像显著性特征的铸坯表面缺陷检测[J].智能系统学报,2012,7(1):75.
WU Jiawei,YAN Jingqi,FANG Zhihong,et al.Defect detection on a steel slab surface based on the characteristics of an image’s saliency region[J].CAAI Transactions on Intelligent Systems,2012,7():75.
[2]逄增治,郑修楠,李金屏.全钢子午线轮胎X光图像的缺陷检测研究现状[J].智能系统学报,2019,14(4):793.[doi:10.11992/tis.201806014]
PANG Zengzhi,ZHENG Xiunan,LI Jinping.Research status of defect detection in X-ray images of all-steel radial tires[J].CAAI Transactions on Intelligent Systems,2019,14():793.[doi:10.11992/tis.201806014]
[3]王凯旋,任福继,倪红军,等.面向电力设备红外图像的温度值识别算法[J].智能系统学报,2022,17(3):617.[doi:10.11992/tis.202105043]
WANG Kaixuan,REN Fuji,NI Hongjun,et al.Temperature value recognition algorithm for the infrared image of power equipment[J].CAAI Transactions on Intelligent Systems,2022,17():617.[doi:10.11992/tis.202105043]
[4]戚银城,耿劭锋,赵振兵,等.基于特征迁移的螺栓图像超分辨率处理方法[J].智能系统学报,2023,18(4):858.[doi:10.11992/tis.202201009]
QI Yincheng,GENG Shaofeng,ZHAO Zhenbing,et al.A method for super resolution processing of bolt image based on feature transfer[J].CAAI Transactions on Intelligent Systems,2023,18():858.[doi:10.11992/tis.202201009]
[5]高海洋,张明川,葛泉波,等.基于点集匹配的缺陷样本图像生成方法[J].智能系统学报,2023,18(5):1030.[doi:10.11992/tis.202209028]
GAO Haiyang,ZHANG Mingchuan,GE Quanbo,et al.Method of defect sample image generation based on point set matching[J].CAAI Transactions on Intelligent Systems,2023,18():1030.[doi:10.11992/tis.202209028]
[6]王延春秋,葛泉波,刘华平.石板材表面缺陷检测的无监督学习方法[J].智能系统学报,2023,18(6):1344.[doi:10.11992/tis.202212006]
WANG Yanchunqiu,GE Quanbo,LIU Huaping.Unsupervised learning method for surface defect detection of slate materials[J].CAAI Transactions on Intelligent Systems,2023,18():1344.[doi:10.11992/tis.202212006]
[7]吴瑞林,葛泉波,刘华平.基于YOLOX的类增量印刷电路板缺陷检测方法[J].智能系统学报,2024,19(4):1061.[doi:10.11992/tis.202309044]
WU Ruilin,GE Quanbo,LIU Huaping.Class-incremental printed circuit board defect detection method based on YOLOX[J].CAAI Transactions on Intelligent Systems,2024,19():1061.[doi:10.11992/tis.202309044]
[8]吴一全,庞雅轩.手机表面缺陷的机器视觉检测方法研究进展[J].智能系统学报,2025,20(1):33.[doi:10.11992/tis.202312036]
WU Yiquan,PANG Yaxuan.Research progress of mobile phone surface defect detection based on machine vision[J].CAAI Transactions on Intelligent Systems,2025,20():33.[doi:10.11992/tis.202312036]
备注/Memo
收稿日期:2023-10-19。
基金项目:山西省回国留学人员科研项目（2021-046）；山西省自然科学基金项目（202103021224056）；山西省科技合作交流专项（202104041101030）.
作者简介:杨奡飞，硕士研究生，主要研究方向为深度学习和缺陷检测。E-mail：yangaofei1998@163.com。;续欣莹，教授，美国圣荷西州立大学访问学者，中国人工智能学会科普工作委员会常务委员、认知系统与信息处理专委会委员，主要研究方向为人工智能、视觉感知与智能控制。主持国家级、省部级和企业等重要项目20 余项，发表学术论文 80 余篇。E-mail：xuxinying@tyut.edu.cn。;刘华平，教授，博士生导师，中国人工智能学会理事、中国人工智能学会认知系统与信息处理专业委员会副主任，主要研究方向为具身感知与学习，吴文俊人工智能科学技术奖获得者。主持国家自然科学基金重点项目2 项，发表学术论文 100 余篇。E-mail：hpliu@tsinghua.edu.cn。
通讯作者:刘华平. E-mail：hpliu@tsinghua.edu.cn
更新日期/Last Update:
2025-01-05