Application of neural networks for control of printed circuit boards using 3D x-ray microtomography data

Cover Page

Cite item

Full Text

Open Access Open Access
Restricted Access Access granted
Restricted Access Subscription or Fee Access

Abstract

The article discusses a method for detecting PCB defects using neural networks. The analysis of various neural network architectures is carried out to identify the most effective. An approach to filtering data simulating the operation of a microtomograph using convolutional autoencoders is also presented. To assess the quality of the proposed approaches, the mean Average Precision (mAP) metric for the YOLOv8 and Faster R-CNN models was used.

Full Text

Restricted Access

About the authors

V. I. Syryamkin

The National Research Tomsk State University, TSU

Author for correspondence.
Email: svi_tsu@mail.ru
Russian Federation, 634050 Tomsk, Lenin str., 36

F. A. Klassen

The National Research Tomsk State University, TSU

Email: svi_tsu@mail.ru
Russian Federation, 634050 Tomsk, Lenin str., 36

A. N. Bertsun

The National Research Tomsk State University, TSU

Email: svi_tsu@mail.ru
Russian Federation, 634050 Tomsk, Lenin str., 36

References

  1. Ivanenko B.P., Klestov S.А., Syryamkin V.I. Comparison of neural network methods of data preprocessing in solving problems of analysis, diagnosis, and classification of defects of radio-electronic equipment // Avtometriya. 2023. V. 59. No. 2. P. 17—24.
  2. Levkevich V.I., Ivanenko B.P., Klestov S.А. Classification of printed circuit board surface defects by neural network methods / Innovation-2023. The XIX International School-Conference of Students, Graduate Students and Young Scientists, Tomsk, Russia. 2023. P. 232—234.
  3. Tsvinger V.А., Syryamkin V.I., Klestov S.А. Application of machine vision in quality control of printing covers / Innovation-2023. The XIX International School-Conference of Students, Graduate Students and Young Scientists, Tomsk, Russia. 2023. P. 225—228.
  4. Alghassab M.A. Defect detection in printed circuit boards with pre-trained feature extraction methodology with convolution neural networks // Computers, Materials & Continua. 2022. V. 70. No. 1. P. 637—652.
  5. Adibhatla V.A., Shieh J.S., Abbod M.F., Chih H.C., Hsu C.C., Cheng J. Detecting Defects in PCB using Deep Learning via Convolution Neural Networks / 2018 13th International Microsystems, Packaging, Assembly and Circuits Technology Conference (IMPACT), Taipei, Taiwan. 2018. P. 202—205.
  6. Chen W., Huang Z., Mu Q., Sun Y. PCB Defect Detection Method Based on Transformer-YOLO // IEEE Access. 2022. V. 10. P. 129480—129489.
  7. Terven Juan, Cordova-Esparza Diana-Margarita, Romero-González Julio-Alejandro. A Comprehensive Review of YOLO Architectures in Computer Vision: From YOLOv1 to YOLOv8 and YOLO-NAS // Machine Learning and Knowledge Extraction. 2023. V. 5. No. 4. P. 1680—1716.
  8. Yujie Yang, Haiyan Kang. An Enhanced Detection Method of PCB Defect Based on Improved YOLOv7 // Electronics. 2023. V. 12. No. 9. P. 2120.
  9. Adibhatla V.A., Chih H.C., Hsu C.C., Cheng J., Abbod M.F., Shieh J.S. Defect Detection in Printed Circuit Boards Using You-Only-Look-Once Convolutional Neural Networks // Electronics. 2020. V. 9. No. 9.
  10. Girshick R. Fast R-CNN / 2015 IEEE International Conference on Computer Vision (ICCV), Santiago, Chile. 2015. P. 1440—1448.
  11. Shaoqing Ren, He Kaiming, Girshick Ross, Sun Jian. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks // IEEE Transactions on Pattern Analysis and Machine Intelligence. 2016. V. 39. No. 6. P. 1137—1149.
  12. Girshick R., Donahue J., Darrell T., Mali J. Rich feature hierarchies for accurate object detection and semantic segmentation / Proceedings of the IEEE conference on computer vision and pattern recognition. 2014. P. 580—587.
  13. Liu L., Ouyang W., Wang X., Fieguth P., Chen J., Liu X., Pietikäinen M. Deep learning for generic object detection: A survey // International journal of computer vision. 2020. V. 128. P. 261—318.
  14. Tang S., He F., Huang X., Yang J. Online PCB defect detector on a new PCB defect dataset // arXiv preprint arXiv:1902.06197. 2019.
  15. Syryamkin V.I., Klestov S.A., Suntsov S.B., Ivanenko B.P., Levkevich V.I. Digital X-ray Tomography // London: Red Square Scientific. 2023. 296 p.
  16. Syryamkin V.I., Klestov S.А., Suncov S.B. Design of 3d x-ray microtomograph based on its «digital twin» // Defectoskopyia. 2022. No. 11. P. 56—65.
  17. Khilchuk М.D., Klestov S.А. X-ray digital tomography for diagnostics of radio-electronic equipment elements / Innovation-2023. The XIX International School-Conference of Students, Graduate Students and Young Scientists, Tomsk, Russia. 2023. P. 229—231.
  18. Minikaev R.R., Klestov S.A. Development of a mathematical model and software for defectoscopy of images of an intelligent x-ray 3d microtomograph based on the neuro-fuzzy method of analysis, diagnosis and classification of defects in radio electronic equipment / Innovation-2022. The XVIII International School-Conference of Students, Graduate Students and Young Scientists Tomsk, Russia. 2022. P. 184—187.
  19. Davis J., Goadrich M. The relationship between Precision-Recall and ROC curves / Proceedings of the 23rd international conference on Machine learning. 2006. P. 233—240.
  20. Venkataraman P. Image denoising using convolutional autoencoder // arXiv preprint arXiv:2207.11771. 2022.
  21. Baldi P. Autoencoders, unsupervised learning, and deep architectures / Proceedings of ICML workshop on unsupervised and transfer learning. JMLR Workshop and Conference Proceedings. 2012. P. 37—49.
  22. Kim J., Ko J., Choi H., Kim H. Printed circuit board defect detection using deep learning via a skip-connected convolutional autoencoder // Sensors. 2021. V. 21. No. 15. P. 4968.
  23. Zhao G., Liu J., Jiang J., Guan H., Wen J.R. Skip-connected deep convolutional autoencoder for restoration of document images / 2018 24th International Conference on Pattern Recognition (ICPR). 2018. P. 2935—2940.
  24. Keras. URL: https://keras.io/
  25. Document № 2024663941. A program for image filtering based on neural network algorithms for the diagnosis of printed circuit boards and REA (PC).

Supplementary files

Supplementary Files
Action
1. JATS XML
Download
2. Fig. 1. Neural network model.

Download (142KB)
Indexing metadata
3. Fig. 2. YOLO architecture, the image is taken from the work "Deep Learning for Generic Object Detection: A Survey" [13].

Download (239KB)
Indexing metadata
4. Fig. 3. Architecture of Faster RCNN, the image is taken from the work "Deep Learning for Generic Object Detection: A Survey" [13].

Download (483KB)
Indexing metadata
5. Fig. 4. Examples of manufactured printed circuit boards.

Download (605KB)
Indexing metadata
6. Fig 5. The number of defects of each type in the presented data set (here the name of the defects is indicated on the abscissa axis, and the total number of defects of a specific type is indicated on the ordinate axis).

Download (152KB)
Indexing metadata
7. Fig. 6. Comparison of the results of the autoencoder: a — original images; b — noisy images; c — images cleared using the autoencoder.

Download (389KB)
Indexing metadata
8. Fig. 7. Block diagram of the software.

Download (230KB)
Indexing metadata

Copyright (c) 2025 Russian Academy of Sciences