TMK Research Center LLC, Moscow, Russia

I. Yu. Pyshmintsev, Dr. Eng., General Director
E. A. Shkuratov, Cand. Eng., Head of Digitalization and Artificial Intelligence Technologies Dept., e-mail: evgeniy.shkuratov@tmk-group.com
D. V. Levshin, Machine Learning Engineer, Digitalization and Artificial Intelligence Technologies Dept.

Seversky Pipe Plant Branch, Polevskoy, Russia
K. P. Pyankov, Deputy Head of Technical Dept.

The paper presents the results of a study aimed at developing and implementing an intelligent system for automated quality control of pierced shells in a tube rolling line. The proposed approach is based on the use of computer vision technologies and machine learning methods for the identification and classification of geometric imperfections on the rear end face of billets formed during the piercing of continuously cast preforms. Particular attention is paid to “crown” and “earring” defects, which have a decisive influence on the geometry and internal surface quality of the finished tubes. A comprehensive analysis of industrial data was carried out, and a representative image dataset was compiled, containing over one thousand samples across three surface condition classes. A convolutional neural network based on the YOLOv11 architecture was employed, providing high detection accuracy under variable lighting and object orientation. Model training was performed using data augmentation and regularization techniques. The resulting model achieved an average classification accuracy of 98.1 %, with F1-scores ranging from 0.95 to 0.98. The developed model was integrated into an industrial interface that performs real-time image processing, automatic defect identification, visualization, and data archiving synchronized with piercing process parameters. The implementation of this system enabled the transition from discrete and subjective visual inspection to continuous, objective quality monitoring. The obtained results confirm the effectiveness of artificial intelligence methods in industrial machine vision tasks and demonstrate their potential for further development within the framework of digital and intelligent manufacturing.

1. Pyshmintsev I. Yu., Shkuratov E. A., Yashin G. A., Kosmin I. V., Rosolenko S. K., Bushin R. O., Pyatkov V. L., Murzin A. V., Izgorev O. P., Mazurin E. V. Development of a hardware and software complex for in-line pipe inspection in continuous production. Metallurg. 2025. No. 6. pp. 82–87. DOI: 10.52351/00260827_2025_6_82
2. Pyshmintsev I. Yu., Shkuratov E. A., Kosmin I. V., Rosolenko S. K., Yashin G. A. Development of an adaptive system for monitoring the geometric parameters of hot-rolled pipes based on machine vision technologies. Chernye Metally. 2025. No. 12. pp. 37–44.
3. Guo D., Zhang C., Yang G., Xue T., Ma W., Liu L., Ren J. Siamese-RCNet: Defect detection model for complex textured surfaces with few annotations. Electronics. 2024. Vol. 13, Iss. 24. 4873. DOI: 10.3390/electronics13244873
4. Curtic-Hodzic M., Ajkunic A., Sokić E., Salihbegović A., Arapovic L., Osmić N., Konjicija S. Automatic leather defect detection and classification using single-channel and multi-channel neural networks. IEEE Access. 2025. DOI: 10.1109/access.2025.3590789
5. Li D., Xie Q., Yu Z., Wu Q., Zhou J., Wang J. Sewer pipe defect detection via deep learning with local and global feature fusion. Automation in Construction. 2021. Vol. 129, Iss. 2. 103823. DOI: 10.1016/j.autcon.2021.103823
6. Yang Z., Zhang M., Li C., Meng Z., Li Y., Chen Y., Liu L. Image сlassification for automobile pipe joints surface defect detection using wavelet decomposition and convolutional neural network. IEEE Access. 2022. Vol. 10, Iss. 2. pp. 77191–77204. DOI: 10.1109/access.2022.3178380
7. Toennies K. D. Image classification: A computer vision task. Springer, 2024. pp. 1–17. DOI: 10.1007/978-981-99-7882-3_1
8. Chernykh I. N., Struin D. O., Shkuratov E. A. Determination of rolling technological factors contributing to the occurrence of surface defects on pipes. Vestnik Yuzhno-Uralskogo gosudarstvennogo universiteta. Seriya: Metallurgiya. 2018. Vol. 18, No. 3. pp. 51–58. DOI: 10.14529/met180306
9. Korsakov A. A., Mikhalkin D. V., Alyutina E. V. et al. Development of measures to reduce defects on the internal surface of oil-grade pipes. Proceedings of the XII Congress of Rollers. 2019. Vol. 1. pp. 102–107.
10. Mikhalkin D. V., Korsakov A. A., Panasenko O. A., Pyankov K. P. Parameters of the deformation zone and boundary conditions of the piercing process. Metallurg. 2021. No. 2. pp. 19–26.
11. Bogatov A. A., Nukhov D. Sh., Toporov V. A., Panasenko O. A. Influence of the plug`s position in the deformation zone on the change in the shape of the billet during piercing. Proizvodstvo prokata. 2017. No. 1. pp. 3–8.
12. Bilan I. T., Trubnikov K. V., Zvonarev D. Yu., Noskova M. N. Detection and classification of rolling defects on the ends of shells using a convolutional neural network. Vestnik Yuzhno-Uralskogo gosudarstvennogo universiteta. Seriya: Metallurgiya. 2023. Vol. 23, No. 1. pp. 22–29. DOI: 10.14529/met230103
13. Goodfellow I., Bengio Y., Courville A. Deep Learning. Cambridge, MA: MIT Press, 2016. Available at: http://www.deeplearningbook.org
14. Cubuk E. D., Zoph B., Mane D., Vasudevan V. K., Le Q. V. AutoAugment: Learning Augmentation Policies from Data. arXiv: Computer Vision and Pattern Recognition. 2018. Available at: https://arxiv.org/abs/1805.09501
15. Shorten C., Khoshgoftaar T. M. A survey on Image data augmentation for deep learning. Journal of Big Data. 2019. Vol. 6. 60. DOI: 10.1186/s40537-019-0197-0
16. Ultralytics. YOLOv11 Documentation. 2024. Available at: https://docs.ultralytics.com/models/yolo11/ (21.10.2025).
17. Khanam R., Hussain M. YOLOv11: An overview of the Key Architectural Enhancements. arXiv, 2024. Available at: https://arxiv.org/abs/2410.17725
18. Loshchilov I., Hutter F. Decoupled weight decay regularization. ICLR, 2019. Available at: https://arxiv.org/abs/1711.05101
19. Kingma D. P., Ba J. Adam: A method for stochastic optimization. arXiv, 2014. Available at: https://arxiv.org/abs/1412.6980
20. Belkin M., Hsu D., Ma S., Mandal S. Reconciling modern machine-learning practice and the classical bias–variance trade-off. Proceedings of the National Academy of Sciences of the United States of America. 2019. Vol. 116, No. 32. pp. 15849–15854. DOI: 10.1073/pnas.1903070116
21. Adeyemi T. Defect detection in manufacturing: an integrated deep learning approach. Journal of Computer and Communications. 2024. Vol. 12. pp. 153–176. DOI: 10.4236/jcc.2024.1210011
22. Redmon J., Farhadi A. YOLOv3: An Incremental Improvement. arXiv, 2018. – URL: https://arxiv.org/abs/1804.02767