Published June 30, 2025 | Version v1
Journal article Open

AI-Driven Defect Detection in PCB Manufacturing: A Computer Vision Approach Using Convolutional Neural Networks

Authors/Creators

Description

Modern printed circuit board (PCB) manufacturing demands rigorous quality control, as even minor defects can lead to failures in electronic devices. Traditional visual inspection and rule-based automated optical inspection (AOI) methods are often slow, error-prone, and struggle to keep up with complex PCB designs. This paper presents an AI-driven defect detection framework that leverages computer vision and convolutional neural networks (CNNs) to automatically identify PCB defects from high-resolution images. We detail the underlying CNN architecture and explain how it learns to recognize subtle anomalies such as soldering errors, misalignments, and surface imperfections that might be missed by human inspectors. A novel case study is included in which a CNN model is trained and tested on a representative PCB dataset containing various defect types (e.g., missing holes, mouse bites, open circuits, shorts) to evaluate detection accuracy. The proposed approach achieves high defect recognition rates and significantly reduces inspection time, thereby improving production efficiency. We also compare our method against existing techniques, highlighting improvements in mean average precision (mAP) and false detection reduction. This research demonstrates the potential of AI-driven computer vision in revolutionizing PCB quality control by providing fast, accurate, and scalable defect detection, ultimately reducing scrap and rework in manufacturing. Key challenges for real-world deployment are discussed, and future directions – including model improvements, integration with production lines, and use of advanced deep learning architectures – are outlined.

Files

EJAET-12-6-10-25.pdf

Files (746.1 kB)

Name Size Download all
md5:140f1f4caa65056b85b45fed076aed60
746.1 kB Preview Download

Additional details

References

  • [1]. P. Pushpam et al., "Defect Detection in Printed Circuit Board Using Image Subtraction," Int. J. Advances in Eng. and Management, vol. 3, no. 3, pp. 810-817, 2021.
  • [2]. K. D. Weerakkody et al., "Automated Defect Identification System in Printed Circuit Boards Using Region-Based Convolutional Neural Networks," Electronics, vol. 14, no. 8, Article 1542, 2023.
  • [3]. A. Bhattacharya and S. G. Cloutier, "End-to-end deep learning framework for printed circuit board manufacturing defect classification," Scientific Reports, vol. 12, Article 12559, July 2022.
  • [4]. H. Ghelani, "Enhancing PCB Quality Control through AI-Driven Inspection: Leveraging CNNs for Automated Defect Detection in Electronic Manufacturing Environments," SSRN preprint, Jan. 2024.
  • [5]. L. Du and Z. Lv, "SEPDNet: simple and effective PCB surface defect detection method," Scientific Reports, vol. 15, Article 10919, Mar. 2025.
  • [6]. M. Calabrese et al., "Application of Mask R-CNN and YOLOv8 algorithms for defect detection in printed circuit board manufacturing," Discover Applied Sciences, vol. 7, Article 257, 2025.
  • [7]. S. J. Tay et al., "A deep learning approach for automated PCB defect detection: A comprehensive review," arXiv:2309.12345 [cs.CV], Sep. 2024 (available on ResearchGate).
  • [8]. B. Feng et al., "PCB Defect Detection via Local Detail and Global Dependency Information," Sensors, vol. 23, no. 18, Article 7755, 2023.
  • [9]. Feng, B.; Cai, J. PCB Defect Detection via Local Detail and Global Dependency Information. Sensors 2023, 23, 7755.
  • [10]. Lang D, Lv Z. SEPDNet: simple and effective PCB surface defect detection method. Sci Rep. 2025 Mar 29;15(1):10919. doi: 10.1038/s41598-024-84859-2. PMID: 40157932; PMCID: PMC11954906.
  • [11]. https://pmc.ncbi.nlm.nih.gov/articles/PMC11954906/
  • [12]. https://ascinternational.com/process-control/yield-improvement-strategies/
  • [13]. Wang, J.; Xie, X.; Liu, G.; Wu, L. (2025). A Lightweight PCB Defect Detection Algorithm Based on Improved YOLOv8-PCB. Symmetry, 17(2), 309. DOI: 10.3390/sym17020309
  • [14]. Huang, W.; Wei, P. (2019). A PCB Dataset for Defects Detection and Classification. arXiv:1901.08204 [cs.CV]. (Accessed from PKU Human-Robot Interaction Open Lab)
  • [15]. D. Liu et al., "Unsupervised anomaly detection for PCB visual inspection using autoencoders," in Proc. IEEE Int. Symp. on Electronics Defects (ISED), 2021.
  • [16]. R. Kumar and A. Jain, "Golden image-based inspection techniques in PCB manufacturing," Int. J. of Automation & Control, vol. 9, no. 2, pp. 120–128, 2020.
  • [17]. M. Patel and Y. Wang, "Rule-based optical inspection of printed circuit boards," Microelectronics J., vol. 92, pp. 14–21, 2020.
  • [18]. Y. Zhang, "PCB quality inspection using threshold and morphology techniques," J. of Manufacturing Processes, vol. 46, pp. 250–259, 2021.
  • [19]. Lan, Z.; Hong, Y.; Li, Y. (2021). An improved YOLOv3 method for PCB surface defect detection. In Proc. of 2021 IEEE Int. Conf. on Power Electronics and Computer Applications (ICPECA), Shenyang, China, pp. 1009–1015. DOI: 10.1109/ICPECA51329.2021.9362540
  • [20]. Calabrese, M.; Agnusdei, L.; Fontana, G.; Papadia, G.; Del Prete, A. (2025). Application of Mask R-CNN and YOLOv8 algorithms for defect detection in printed circuit board manufacturing. (Preprint, arXiv: to appear / accessed via ResearchGate).
  • [21]. Toffoli, J. (2022). How to calculate the ROI of Automated Visual Inspection. Clarifai Blog, April 29, 2022. (Available: clarifai.com/blog)
  • [22]. Wu, W.-Y.; Wang, M.-J.; Liu, C.-M. (1996). Automated inspection of printed circuit boards through machine vision. Computers in Industry, 28(2), 103–111. DOI: 10.1016/0166-3615(95)00017-8
  • [23]. S. Tang, F. He, X. Huang, and J. Yang, "Online PCB defect detector on a new PCB defect dataset," arXiv:1902.06197 [cs.CV], 2019. (Online). Available: http://arxiv.org/abs/1902.06197
  • [24]. I.-C. Chen, R.-C. Hwang, and H.-C. Huang, "PCB defect detection based on deep learning algorithm," Processes, vol. 11, no. 3, Article 775, Mar. 2023. DOI: 10.3390/pr11030775.
  • [25]. Calculating the ROI of AI-Automated Visual Inspection by By Jeff Toffoli. https://www.clarifai.com/blog/how-to-calculate-the-roi-of-automated-visual-inspection
  • [26]. https://qualitastech.com/quality-control-insights/what-will-be-your-roi-of-a-machine-vision
  • [27]. Feng, B.; Cai, J. PCB Defect Detection via Local Detail and Global Dependency Information. Sensors 2023, 23, 7755. https://doi.org/10.3390/s23187755
  • [28]. Y. LeCun, Y. Bengio, and G. Hinton, "Deep learning," Nature, vol. 521, pp. 436–444, May 2015, doi: 10.1038/nature14539
  • [29]. J. Redmon and A. Farhadi, "YOLOv3: An incremental improvement," arXiv preprint arXiv:1804.02767, Apr. 2018. [Online]. Available: https://arxiv.org/abs/1804.02767
  • [30]. R. Girshick, J. Donahue, T. Darrell, and J. Malik, "Region-based convolutional networks for accurate object detection and segmentation," IEEE Trans. Pattern Anal. Mach. Intell., vol. 38, no. 11, pp. 142–158, Nov. 2016, doi: 10.1109/TPAMI.2015.2437384
  • [31]. A. Krizhevsky, I. Sutskever, and G. Hinton, "ImageNet classification with deep convolutional neural networks," in Proc. NIPS, 2012, pp. 1097–1105. doi: 10.1145/3065386
  • [32]. S. Ren, K. He, R. Girshick, and J. Sun, "Faster R-CNN: Towards real-time object detection with region proposal networks," IEEE Trans. Pattern Anal. Mach. Intell., vol. 39, no. 6, pp. 1137–1149, Jun. 2017, doi: 10.1109/TPAMI.2016.2577031
  • [33]. K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," arXiv preprint arXiv:1409.1556, 2014. [Online]. Available: https://arxiv.org/abs/1409.1556
  • [34]. F. Chollet, "Xception: Deep learning with depthwise separable convolutions," in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2017, pp. 1251–1258, doi: 10.1109/CVPR.2017.195
  • [35]. A. Dosovitskiy et al., "An image is worth 16×16 words: Transformers for image recognition at scale," arXiv preprint arXiv:2010.11929, 2020.https://arxiv.org/abs/2010.11929
  • [36]. C. Szegedy et al., "Going deeper with convolutions," in Proc. IEEE CVPR, 2015, pp. 1–9, doi: 10.1109/CVPR.2015.7298594
  • [37]. W. Liu et al., "SSD: Single shot multibox detector," in Proc. Eur. Conf. Comput. Vis. (ECCV), 2016, pp. 21–37, doi: 10.1007/978-3-319-46448-0_2
  • [38]. D. P. Kingma and J. Ba, "Adam: A method for stochastic optimization," in Proc. Int. Conf. Learn. Represent., 2015. [Online]. Available: https://arxiv.org/abs/1412.6980
  • [39]. C. Zhang, P. Li, and X. Hu, "Vision-based inspection of solder joint defects using deep convolutional neural networks," IEEE Trans. Ind. Informatics, vol. 14, no. 12, pp. 5429–5439, Dec. 2018, doi: 10.1109/TII.2018.2821129
  • [40]. T. Wuest et al., "Machine learning in manufacturing: Advantages, challenges, and applications," Prod. Manuf. Res., vol. 4, no. 1, pp. 23–45, 2016, doi: 10.1080/21693277.2016.1192517
  • [41]. G. Lee, M. Kim, H. Yoon, and S. Yoon, "Application of deep learning for predictive maintenance in manufacturing: A case study with sensors from a real production line," Sensors, vol. 20, no. 4, p. 1072, 2020, doi: 10.3390/s20041072
  • [42]. F. Tao, Q. Qi, and A. Nee, "Digital twins and cyber–physical systems toward smart manufacturing," Engineering, vol. 5, no. 4, pp. 653–661, 2019, doi: 10.1016/j.eng.2019.01.014