RESEARCH OF PUBLIC ARTIFICIAL INTELLIGENCE DATASETS FOR RAILWAY APPLICATIONS
https://doi.org/10.53360/2788-7995-2026-3(23)-22
Abstract
The development of artificial intelligence (AI) technologies in the railway industry directly depends on the availability of high-quality and representative public datasets. This article provides a systematic analysis of open datasets used for the tasks of rail defect detection, contact network diagnostics, rolling stock monitoring and railway infrastructure safety. The analysis was conducted following the PRISMA methodology: the search covered IEEE Xplore, Scopus and Web of Science, as well as the Kaggle, GitHub and Zenodo repositories. In total, 30 key publications and 13 open datasets were analysed and classified by sensor type (RGB, LiDAR, vibration, multi-sensor), machine learning task (detection, segmentation, classification), data volume and annotation format. The analysis shows that most datasets are based on RGB images, while the share of multimodal datasets is gradually increasing. Key limitations of existing datasets were identified -small sample size, class imbalance, limited geographical variability and the absence of unified benchmark protocols and directions for further research were defined. The results highlight the need to create large multimodal datasets and to develop unified standards for the annotation and evaluation of AI models.
About the Authors
G. K. AbishevaKazakhstan
Gulsipat Abisheva – doctoral student of the Department of artificial intelligence
10000 Kazakhstan, Astana city, 2 Satbaev st.
B. Sh. Razakhova
Kazakhstan
Bibigul Razakhova – с.t.s., Head of the Department of Artificial Intelligence Technologies
10000 Kazakhstan, Astana city, 2 Satbaev st.
T. A. Aidynov
Kazakhstan
Tolegen Aidynov – doctoral student of the Department of Information Security
10000 Kazakhstan, Astana city, 2 Satbaev st.
M. K. Saparbek
Kazakhstan
Minira Saparbek – master of the Department of artificial intelligence
10000 Kazakhstan, Astana city, 2 Satbaev st.
References
1. European Union Agency for Railways. Railway Safety Performance in the European Union 2022. – Luxembourg: Publications Office of the European Union, 2022.
2. Papaelias M. A review on non-destructive evaluation of rails / M. Papaelias, C. Roberts, C. Davis // Insight -Non-Destructive Testing and Condition Monitoring. – 2008. – Vol. 50, № 7. – P. 367-373.
3. Goodfellow I. Deep Learning / I. Goodfellow, Y. Bengio, A. Courville. – Cambridge: MIT Press, 2016. – 775 p.
4. Rail surface defect detection method based on YOLOv3 deep learning networks / S. Yanan et al // 2018 Chinese Automation Congress (CAC). – IEEE, 2018. – P. 1563-1568. https://doi.org/10.1109/CAC.2018.8623082.
5. Ning S. Research on the method of foreign object detection for railway tracks based on deep learning / S. Ning, F. Ding, B. Chen // Sensors. – 2024. – Vol. 24, № 14. – Article 4483. https://doi.org/10.3390/s24144483.
6. Deep learning approach for automated railroad ballast condition evaluation / J. Luo et al // Proceedings of the 5th International Conference on Transportation Geotechnics (ICTG 2024). Lecture Notes in Civil Engineering. – Springer, 2025. – Vol. 402. – P. 49-57. https://doi.org/10.1007/978-981-97-8213-0_7.
7. A robust pantograph-catenary interaction condition monitoring method based on deep convolutional network / D. Zhang et al // IEEE Transactions on Instrumentation and Measurement. – 2020. – Vol. 69, № 5. – P. 1920-1929. https://doi.org/10.1109/TIM.2019.2920721.
8. Applications of machine learning to machine fault diagnosis: A review and roadmap / Y. Lei et al // Mechanical Systems and Signal Processing. – 2020. – Vol. 138. – Article 106587. https://doi.org/10.1016/j.ymssp.2019.106587.
9. Deep convolutional neural networks for detection of rail surface defects / S. Faghih-Roohi et al // Proceedings of the 2016 International Joint Conference on Neural Networks (IJCNN). – IEEE, 2016. – P. 2584-2589. https://doi.org/10.1109/IJCNN.2016.7727522.
10. MCnet: Multiple context information segmentation network of no-service rail surface defects / D. Zhang et al // IEEE Transactions on Instrumentation and Measurement. – 2021. – Vol. 70. – Article 5004909. https://doi.org/10.1109/TIM.2020.3040890.
11. Deep residual learning for image recognition / K. He et al // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. – 2016. – P. 770-778.
12. Ronneberger O. U-Net: Convolutional networks for biomedical image segmentation / O. Ronneberger, P. Fischer, T. Brox // Medical Image Computing and Computer-Assisted Intervention. – Cham: Springer, 2015. – P. 234-241.
13. You Only Look Once: Unified, real-time object detection / J. Redmon et al // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. – 2016. – P. 779-788.
14. Faster R-CNN: Towards real-time object detection with region proposal networks / S. Ren et al // Advances in Neural Information Processing Systems. – 2015. – Vol. 28.
15. Encoder-decoder with atrous separable convolution for semantic image segmentation / L.-C. Chen et al // Proceedings of the European Conference on Computer Vision. – 2018. – P. 801-818.
16. An image is worth 16x16 words: Transformers for image recognition at scale / A. Dosovitskiy et al // International Conference on Learning Representations. – 2021.
17. RailSem19: A Dataset for Semantic Rail Scene Understanding / O. Zendel et al // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. – 2019.
18. OSDaR23: Open sensor data for rail 2023 / R. Tagiew et al // 2023 8th International Conference on Robotics and Automation Engineering. – IEEE, 2023. – P. 270-276.
19. A systematic review of artificial intelligence public datasets for railway applications / M.J. Pappaterra et al // Infrastructures. – 2021. – Vol. 6, № 9. – Article 136.
20. A survey on audio-video based defect detection through deep learning in railway maintenance / L. De Donato et al // IEEE Access. – 2022. – Vol. 10. – P. 65376-65400. https://doi.org/10.1109/ACCESS.2022.3183102.
21. Liu S. A review of applications of visual inspection technology based on image processing in the railway industry / S. Liu, Q. Wang, Y. Luo // Transportation Safety and Environment. – 2019. – Vol. 1, № 3. – P. 185-204. https://doi.org/10.1093/tse/tdz007.
22. Artificial intelligence in railway transport: Taxonomy, regulations and applications / N. Bešinović et al // IEEE Transactions on Intelligent Transportation Systems. – 2022. – Vol. 23, № 9. – P. 14011-14024. https://doi.org/10.1109/TITS.2021.3131637.
23. ImageNet: A large-scale hierarchical image database / J. Deng et al // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. – 2009. – P. 248-255.
24. Microsoft COCO: Common objects in context / T.-Y. Lin et al // Proceedings of the European Conference on Computer Vision. – Cham: Springer, 2014. – P. 740-755.
25. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews / M. Page et al // BMJ. – 2021. – Vol. 372. – Article n71.
26. Rail-5k: A real-world dataset for rail surface defects detection / Z. Zhang et al // arXiv preprint. – 2021. – arXiv:2106.14366.
27. Guo Q. SynRailObs: A synthetic dataset for obstacle detection in railway scenarios / Q. Guo, J. Rambach // arXiv preprint. – 2025. -arXiv:2505.10784.
28. OSDaR-AR: Enhancing Railway Perception Datasets via Multi-modal Augmented Reality / F. Nesti et al // arXiv preprint. – 2026. – arXiv:2602.22920.
29. Unsupervised saliency detection of rail surface defects using stereoscopic images / M. Niu et al // IEEE Transactions on Industrial Informatics. – 2021. – Vol. 17, № 3. – P. 2271-2281.
30. WHU-Railway3D: A diverse dataset and benchmark for railway point cloud semantic segmentation / B. Qiu et al // IEEE Transactions on Intelligent Transportation Systems. – 2024. – Vol. 25. – P. 20900-20916.
31. FRSign: A large-scale traffic light dataset for autonomous trains / J. Harb et al // arXiv preprint. – 2020.
32. Leibner P. GERALD: A novel dataset for the detection of German mainline railway signals / P. Leibner, F. Hampel, C. Schindler // Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit. – 2023.
33. RailGoerl24: Görlitz Rail Test Center CV Dataset 2024 / R. Tagiew et al // arXiv preprint. – 2025. – arXiv:2504.00204.
34. SynDRA: Synthetic dataset for railway applications / G. D'Amico et al // 2025 IEEE/CVF Winter Conference on Applications of Computer Vision. – IEEE, 2025. – P. 3437-3446.
35. Eunus S.I. Railway Track Fault Detection [Electronic resource] / S.I. Eunus, S. Hossain, A.E.M. Ridwan // Kaggle Dataset. – 2022. https://www.kaggle.com/datasets/salmaneunus/railway-trackfault-detection (accessed: 09.07.2026).
36. Augmented dataset for vision-based analysis of railroad ballast via multi-dimensional data synthesis / K. Ding et al // Algorithms. – 2024. – Vol. 17, № 8. – Article 367.
Review
For citations:
Abisheva G.K., Razakhova B.Sh., Aidynov T.A., Saparbek M.K. RESEARCH OF PUBLIC ARTIFICIAL INTELLIGENCE DATASETS FOR RAILWAY APPLICATIONS. Bulletin of Shakarim University. Technical Sciences. 2026;(3(23)):216-227. (In Russ.) https://doi.org/10.53360/2788-7995-2026-3(23)-22
JATS XML















