Investigation of Rail Defects Induced by Rolling Contact Fatigue Using YOLOv8 - Based Computer Vision and Spatial - Temporal Post - Processing
DOI:
https://doi.org/10.5281/zenodo.23363204Keywords:
Rail Crack Detection, YOLOv8, Video Analytics, Temporal Filtering, Railway Infrastructure Monitoring, Finite Element AnalysisAbstract
Rail surface defects caused by rolling contact fatigue (RCF) significantly affect railway operational safety and infrastructure reliability. This study proposes an automated rail surface defect detection framework based on the YOLOv8 object detection architecture combined with a lightweight spatial-temporal post-processing module designed for video-based monitoring. We collected a dataset under real operating conditions in Kazakhstan, including snow, variable illumination, and motion-induced blur, with temperatures down to -32 °C. Four object classes were annotated: rail, crack defect, joint, and bolt_joint. The baseline YOLOv8 model achieved an average mAP@0.5 of 0.849. However, crack detection recall was limited (0.644) due to small defect size and temporal instability in video sequences. The proposed spatial–temporal confirmation mechanism increased crack recall to 0.688 without modifying the neural network architecture or increasing inference complexity. Additionally, finite element simulations of R65 rails with longitudinal cracks (2 mm and 4 mm depth) under a 120 kN contact load demonstrated stress concentration growth and a reduction in safety factor from 3.48 (intact rail) to 3.15 (4 mm crack). The results confirm the engineering relevance of early crack detection and demonstrate the feasibility of integrating YOLO-based monitoring with structural integrity assessment.
References
Abdurasulov, S., Zayniddinov, N., Khamidov, O., Yusufov, A., & Jamilov, S. (2025). Stress-strain state analysis of the crossbeam of the main frame of industrial electric locomotives PE2M and PE2U. In AIP Conference Proceedings (Vol. 3256, pp. 060011). AIP Publishing LLC. https://doi.org/10.1063/5.0266927
Alimov, M., & Meiramkhanov, T. (2024). Domain generalization in autonomous driving: Evaluating yolov8s, rt-detr, and yolo-nas with the road-almaty dataset. arXiv preprint arXiv:2412.12349. https://doi.org/10.48550/arXiv.2412.12349
Bai, T., Gao, J., Yang, J., & Yao, D. (2021). A study on railway surface defects detection based on machine vision. Entropy, 23(11), 1437. https://doi.org/10.3390/e23111437
Barlybayev, A., Amangeldy, N., Kurmetbek, B., Krak, I., Razakhova, B., Tursynova, N., & Turebayeva, R. (2024). Personal protective equipment detection using YOLOv8 architecture on object detection benchmark datasets: a comparative study. Cogent Engineering, 11(1), 2333209. https://doi.org/10.1080/23311916.2024.2333209
Baur, J., & Nitsche, F. O. (2025). A False-Positive-Centric Framework for Object Detection Disambiguation. Remote Sensing, 17(14), 2429. https://doi.org/10.3390/rs17142429
Bewley, A., Ge, Z., Ott, L., Ramos, F., & Upcroft, B. (2016). Simple online and realtime tracking. In 2016 IEEE international conference on image processing (ICIP) (pp. 3464-3468). Ieee. https://doi.org/10.1109/ICIP.2016.7533003
Bochkovskiy, A., Wang, C.-Y., & Liao, H.-Y. M. (2020). Yolov4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934. https://doi.org/10.48550/arXiv.2004.10934
CanYang, X., Yingying, L., Yongqiang, L., Runliang, T., & Tao, G. (2025). Lightweight rail surface defect detection algorithm based on an improved YOLOv8. Measurement, 242, 115922. https://doi.org/10.1016/j.measurement.2024.115922
Chen, L., Sun, Q., Han, Z., & Zhai, F. (2025). DP-YOLO: a lightweight real-time detection algorithm for rail fastener defects. Sensors, 25(7), 2139. https://doi.org/10.3390/s25072139
Du, J., Zhang, R., Gao, R., Nan, L., & Bao, Y. (2024). RSDNet: A new multiscale rail surface defect detection model. Sensors, 24(11), 3579. https://doi.org/10.3390/s24113579
Dumitriu, M., & Gheţi, M. (2018). Influence of the interference of bounce and pitch vibrations upon the dynamic behaviour in the bogie of a railway vehicle. In IOP Conference Series: Materials Science and Engineering (Vol. 400, pp. 042020). IOP Publishing. https://doi.org/10.1088/1757-899X/400/4/042020
Fang, Z., Li, L., Peng, L., Zheng, S., Zhong, Q., & Zhu, T. (2024). YOLOv8n-RSDD: a high-performance low-complexity rail surface defect detection network. Ieee Access, 12, 196249-196265. https://doi.org/10.1109/ACCESS.2024.3466559
Ghodrati, M., Ahmadian, M., & Mirzaeifar, R. (2018). Modeling of rolling contact fatigue in rails at the microstructural level. Wear, 406, 205-217. https://doi.org/10.1016/j.wear.2018.04.016
Gong, W., Akbar, M. F., Jawad, G. N., Mohamed, M. F. P., & Wahab, M. N. A. (2022). Nondestructive testing technologies for rail inspection: A review. Coatings, 12(11), 1790. https://doi.org/10.3390/coatings12111790
Guan, Z., Wang, Z., Zhang, G., Li, L., Zhang, M., Shi, Z., & Jiang, N. (2025). Multi-object tracking review: retrospective and emerging trend. Artificial Intelligence Review, 58(8), 235. https://doi.org/10.1007/s10462-025-11212-y
Huang, Z., Wang, J., Fu, X., Yu, T., Guo, Y., & Wang, R. (2020). DC-SPP-YOLO: Dense connection and spatial pyramid pooling based YOLO for object detection. Information Sciences, 522, 241-258. https://doi.org/10.1016/j.ins.2020.02.067
Izvoltova, J., Izvolt, L., & Sestakova, J. (2022). Analysis of methods used to diagnostics of railway lines. IntechOpen. https://doi.org/10.5772/intechopen.100835
Jamali, M., Davidsson, P., Khoshkangini, R., Ljungqvist, M. G., & Mihailescu, R.-C. (2025). Context in object detection: a systematic literature review. Artificial Intelligence Review, 58(6), 175. https://doi.org/10.1007/s10462-025-11186-x
Jamshiddoust, A. (2025). Improving dynamic stability and strength investigation of a locomotive bolster-type bogie frame through fabricated design and damper implementation. Advances in Mechanical Engineering, 17(12), 16878132251401390. https://doi.org/10.1177/16878132251401390
Kang, D., Emmons, J., Abuzaid, F., Bailis, P., & Zaharia, M. (2017). Noscope: optimizing neural network queries over video at scale. arXiv preprint arXiv:1703.02529. https://doi.org/10.48550/arXiv.1703.02529
Khromova, G., Mukhamedova, Z., & Yutkina, I. (2017). Mathematical model of oscillations of bearing body frame of emergency and repair railcars. Transport problems, 12(1), 93-102. https://doi.org/10.20858/tp.2017.12.1.9
Khromova, G., Radjibaev, D., Zabiyeva, A., Kenesbek, A., & Mavlanov, A. (2025). Calculation of the main parameters of the two-line helical traction transmission of an electric locomotive based on diagnostic parameters. Applied Sciences, 15(4), 1730. https://doi.org/10.3390/app15041730
Liu, Y., Hu, T., Wang, H., Zhuang, J., & Sheng, J. (2024). Dynamic implementation and verification of the gauge changing function of high-speed variable gauge bogies. Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 238(24), 11366-11378. https://doi.org/10.1177/09544062241280688
Luo, H., Cai, L., & Li, C. (2023). Rail surface defect detection based on an improved YOLOv5s. Applied Sciences, 13(12), 7330. https://doi.org/10.3390/app13127330
Lv, D., Meng, J., Meng, G., & Shen, Y. (2025). Railway Fastener Defect Detection Model Based on Dual Attention and MobileNetv3. World Electric Vehicle Journal, 16(9), 513. https://doi.org/10.3390/wevj16090513
Mao, Y., Zheng, S., Li, L., Shi, R., & An, X. (2024). Research on rail surface defect detection based on improved CenterNet. Electronics, 13(17), 3580. https://doi.org/10.3390/electronics13173580
Mi, Z., Chen, R., & Zhao, S. (2023). Research on steel rail surface defects detection based on improved YOLOv4 network. Frontiers in neurorobotics, 17, 1119896. https://doi.org/10.3389/fnbot.2023.1119896
Nezhad, M. S., Larsson, F., Kabo, E., & Ekberg, A. (2024). Finite element analyses of rail head cracks: Predicting direction and rate of rolling contact fatigue crack growth. Engineering Fracture Mechanics, 310, 110503. https://doi.org/10.1016/j.engfracmech.2024.110503
Ning, Y., Wang, L., Qiu, S., Zhang, X., Xing, B., Fu, X., Li, R., Zhang, J., Deng, K., & Wang, H. (2025). Shear strain-assisted cross-twin orientation regulation enhances the plasticity of ZM21 magnesium alloy. Journal of Alloys and Compounds, 183040. https://doi.org/10.1016/j.jallcom.2025.183040
Peng, Z., Wang, C., Ma, Z., & Liu, H. (2019). A multifeature hierarchical locating algorithm for hexagon nut of railway fasteners. IEEE Transactions on Instrumentation and Measurement, 69(3), 693-699. https://doi.org/10.1109/TIM.2019.2905309
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 779-788). https://doi.org/10.1109/CVPR.2016.91
Redmon, J., & Farhadi, A. (2018). Yolov3: An incremental improvement. arXiv preprint arXiv:1804.02767. https://doi.org/10.48550/arXiv.1804.02767
Tarkasis, K., Kaparis, K., & Georgiou, A. C. (2025). Enhancing trustworthiness in real time single object detection. Information Systems Frontiers, 1-12. https://doi.org/10.1007/s10796-025-10584-4
Togizbayeva, B., Zabiyevа, A., Balabekova, K., & Kenesbek, A. (2023). Development of a Calculation Method for Determining the Permissible Dimensions of the Defect 11. Acta Polytechnica Hungarica, 20(4). https://tinyurl.com/mpj8258a
Trollé, B., Baietto, M.-C., Gravouil, A., Mai, S. H., & Prabel, B. (2014). 2D fatigue crack propagation in rails taking into account actual plastic stresses. Engineering Fracture Mechanics, 123, 163-181. https://doi.org/10.1016/j.engfracmech.2014.03.020
Tulekov, A., Togizbayeva, B., Kenesbek, I., Kenesbek, A., & Zabiyeva, A. (2023). The Use of Composite Materials in the Production of Tower Cranes. https://doi.org/10.12700/APH.20.9.2023.9.16
Wang, J., Guo, X., Jing, L., & WANG, K. (2022). Finite element simulations of wheel-rail impact response induced by wheel tread spalling of high-speed trains. Explosion and Shock Waves, 42(4), 045103-045101-045103-045115. https://www.sciengine.com/ESW/doi/10.11883/bzycj-2021-0374
Wang, L., Zhang, Q., & Wang, J. (2023). Microstructural effects on overall dynamics of composites: an analytical method via spatiotemporal nonlocal model: L. Wang et al. Archive of Applied Mechanics, 93(1), 73-91. https://doi.org/10.1007/s00419-022-02206-1
Wang, M., Su, H., Yan, Z., Chen, C., Zheng, C., Lv, B., & Zhang, F. (2026). A Review of Experimental and Simulation Methods for Rolling Contact Fatigue. Metals, 16(2), 168. https://doi.org/10.3390/met16020168
Wang, Y., Zhang, K., Wang, L., & Wu, L. (2024). An improved YOLOv8 algorithm for rail surface defect detection. Ieee Access, 12, 44984-44997. https://doi.org/10.1109/ACCESS.2024.3380009
Wojke, N., Bewley, A., & Paulus, D. (2017). Simple online and realtime tracking with a deep association metric. In 2017 IEEE international conference on image processing (ICIP) (pp. 3645-3649). IEEE. https://doi.org/10.1109/ICIP.2017.8296962
Wu, X., & Yu, S. (2025). Research on the lightweight detection method of rail internal damage based on improved YOLOv8. Journal of Engineering and Applied Science, 72(1), 14. https://doi.org/10.1186/s44147-025-00584-1
Xing, Z., Zhang, Z., Yao, X., Qin, Y., & Jia, L. (2022). Rail wheel tread defect detection using improved YOLOv3. Measurement, 203, 111959. https://doi.org/10.1016/j.measurement.2022.111959
Zhang, C., Xu, D., Zhang, L., & Deng, W. (2023). Rail surface defect detection based on image enhancement and improved YOLOX. Electronics, 12(12), 2672. https://doi.org/10.3390/electronics12122672
Zhexebay, D., Skabylov, A., Turlykozhayeva, D., Bolysbay, A., Turmaganbet, U., Namazbayev, T., & Mingliang, T. (2025). Real-time small object detection with YOLOv8n/8s and YOLOv11n/11s models in complex natural landscapes. Physical Sciences and Technology, 12(1-2), 142-153. https://doi.org/10.26577/phst202512114
Zhuang, L., Qi, H., & Zhang, Z. (2021). The automatic rail surface multi-flaw identification based on a deep learning powered framework. IEEE Transactions on Intelligent Transportation Systems, 23(8), 12133-12143. https://doi.org/10.1109/TITS.2021.3109949
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Reports in Mechanical Engineering
This work is licensed under a Creative Commons Attribution 4.0 International License.