Autonomous Rail Surface Defect Identification Based on an Improved One-Stage Object Detection Algorithm
Publication: Journal of Performance of Constructed Facilities
Volume 38, Issue 5
Abstract
The rail is an important component of track infrastructure, which withstands repeated wheel loading directly, and its condition is related to the safety of train operation. Thus, accurately identifying the size and location of surface defects in rails helps to optimize maintenance strategies, including adjusting regular monitoring and conducting timely repairs. This approach not only mitigates risks but also enhances work efficiency, which has real economic value and brings safety guarantees. This paper aims to build a data-driven model for rail surface defect identification using photos taken in real lines. Two modules, multidirection rectangular convolution (MRC) and cross-scale (CS) feature extraction, are proposed. The results indicate that the detection and classification of multiple rail surface defects can be automated simultaneously with greater accuracy. Among the defects, spalling sees the most significant boost, and its average detection precision increases from 44.1% to 67%. Moreover, the accuracy for detecting a bright contact band and corrugation exceeds 90%, with 0.995 and 0.915, respectively. Compared with the original You Only Look Once algorithm version 8, the mean of average precision (mAP) of the improved network increases from 85.3% to 88.1% when both models are trained for 300 epochs. Additionally, the precise location and size information of the rail surface defects are obtained through postprocessing, providing support for further intelligent track maintenance.
Get full access to this article
View all available purchase options and get full access to this article.
Data Availability Statement
Some or all data, models, or code that support the findings of this study are available from the corresponding author upon reasonable request.
Acknowledgments
The authors acknowledge Projects of Scientific and Technological Research and Development Program of China Railway (L2021G002), National Natural Science Foundation of China (51678445 and 51878661), Key Projects of Shanghai Science and Technology Commission (20dz1203100), fundamental research funds for the central universities (2022-5-ZD-04), and the Department of Transportation Science and Technology Plan Project of Zhejiang Province (Project No. 2023024).
References
Afridi, A., H. Zhu, E. Camacho, G. Deng, and H. Li. 2023. “Numerical modeling of rolling contact fatigue cracks in the railhead.” Eng. Fail. Anal. 143 (Mar): 106838. https://doi.org/10.1016/j.engfailanal.2022.106838.
Al-Juboori, A., H. Zhu, H. Li, J. McLeod, S. Pannila, and J. Barnes. 2023. “Microstructural investigation on a rail fracture failure associated with squat defects.” Eng. Fail. Anal. 151 (Mar): 107411. https://doi.org/10.1016/j.engfailanal.2023.107411.
Gao, R., and S. Fan. 2020. “Research on the propagation characteristics of fatigue cracks on rail surfaces.” Int. J. Appl. Mech. 12 (10): 2050121. https://doi.org/10.1142/S1758825120501215.
Gao, S., M. Cheng, K. Zhao, X. Zhang, M. Yang, and P. Torr. 2021. “Res2Net: A new multi-scale backbone architecture.” IEEE Trans. Pattern Anal. Mach. Intell. 43 (2): 652–662. https://doi.org/10.1109/TPAMI.2019.2938758.
Girshick, R., J. Donahue, T. Darrell, and J. Malik. 2014. “Rich feature hierarchies for accurate object detection and semantic segmentation.” In Proc., 2014 IEEE Conf. On Computer vision and Pattern Recognition, 580. New York: IEEE.
He, D., Z. Zou, Y. Chen, B. Liu, and J. Miao. 2021. “Rail transit obstacle detection based on improved CNN.” IEEE Trans. Instrum. Meas. 70 (Sep): 2515114. https://doi.org/10.1109/TIM.2021.3116315.
He, K., X. Zhang, S. Ren, and J. Sun. 2016. “Identity mappings in deep residual networks.” In Proc., Computer Vision–ECCV 2016: 14th European Conf., 630–645. Berlin: Springer.
Hsieh, C., T. Hsu, and W. Huang. 2022. “An online rail track fastener classification system based on YOLO models.” Sensors 22 (24): 9970. https://doi.org/10.3390/s22249970.
Ignesti, M., A. Innocenti, L. Marini, E. Meli, and A. Rindi. 2014. “Development of a model for the simultaneous analysis of wheel and rail wear in railway systems.” Multibody Syst. Dyn. 31 (2): 191–240. https://doi.org/10.1007/s11044-013-9360-0.
Jiang, J., L. Liu, Y. Cui, and Y. Zhao. 2023. “A nested unet based on multi-scale feature extraction for mixed gaussian-impulse removal.” Appl. Sci. 13 (17): 9520. https://doi.org/10.3390/app13179520.
Jin, X., and Q. Liu. 2004. Tribology of wheel and rail. Beijing: China Railway Publishing House.
Jun, Y., H. Shin, T. Eo, and D. Hwang. 2021. “Joint deep model-based MR image and coil sensitivity reconstruction network (joint-icnet) for fast MRI.” In Proc., 2021 IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR), 5266–5275. New York: IEEE.
Li, C., Q. Guan, S. Xu, Z. Wen, and G. Tao. 2023a. “Analysis on formation mechanism of double contact bands on metro rails.” J. Cent. South Univ. Sci. Technol. 54 (4): 1633–1643.
Li, D., R. Yao, C. Yang, C. Zhao, and L. Zhang. 2023b. “An improved Res2Net-based model for classifying the appearance of deer antler slices.” IEEE Access 11 (Dec): 99705–99715. https://doi.org/10.1109/ACCESS.2023.3290026.
Li, G., Z. Liu, and H. Ling. 2020. “ICNet: Information conversion network for RGB-D based salient object detection.” IEEE Trans. Image Process. 29 (Mar): 4873–4884. https://doi.org/10.1109/TIP.2020.2976689.
Li, P., J. Zheng, P. Li, H. Long, M. Li, and L. Gao. 2023c. “Tomato maturity detection and counting model based on MHSA-YOLOv8.” Sensors 23 (15): 6701. https://doi.org/10.3390/s23156701.
Li, X., T. Yang, J. Zhang, Y. Cao, Z. Wen, and X. Jin. 2016. “Rail wear on the curve of a heavy haul line-numerical simulations and comparison with field measurements.” Wear 366 (Mar): 131–138. https://doi.org/10.1016/j.wear.2016.06.024.
Liu, C., D. Li, and P. Huang. 2021. “ISE-YOLO: Improved squeeze-and-excitation attention module based YOLO for blood cells detection.” In Proc., 2021 IEEE Int. Conf. on Big Data (Big Data), 3911–3916. New York: IEEE.
Machado, M., P. Moreira, P. Flores, and H. Lankarani. 2012. “Compliant contact force models in multibody dynamics: Evolution of the hertz contact theory.” Mech. Mach. Theory 53 (Dec): 99–121. https://doi.org/10.1016/j.mechmachtheory.2012.02.010.
Meng, S., S. Kuang, Z. Ma, and Y. Wu. 2022. “MtlrNet: An effective deep multitask learning architecture for rail crack detection.” IEEE Trans. Instrum. Meas. 71 (Jun): 1–10. https://doi.org/10.1109/TIM.2022.3181940.
Naseri, R., S. Mohammadzadeh, and D. Rizos. 2024. “Rail surface spot irregularity effects in vehicle-track interaction simulations of train-track-bridge interaction.” J. Vib. Control 10775463241232024. https://doi.org/10.1177/10775463241232024.
Ni, Y., J. Mao, H. Wang, Z. Xi, and Y. Xu. 2023. “Toward high-precision crack detection in concrete bridges using deep learning.” J. Perform. Constr. Facil. 37 (3); 04023017. https://doi.org/10.1061/JPCFEV.CFENG-4275.
Redmon, J., and A. Farhadi. 2018. “YOLOv3: An incremental improvement.” Preprint, submitted April 8, 2018. http://arxiv.org/abs/1804.02767.
Ren, S., K. He, R. Girshick, and J. Sun. 2016. “Faster R-CNN: Towards real-time object detection with region proposal networks.” Preprint, submitted April 8, 2016. http://arxiv.org/abs/1506.01497.
Shafiq, M., and Z. Gu. 2022. “Deep residual learning for image recognition: A survey.” Appl. Sci. 12 (18): 8972. https://doi.org/10.3390/app12188972.
Wang, C., A. Bochkovskiy, and H. Liao. 2023. “YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors.” In Proc., IEEE/CVF Conf. on Computer Vision and Pattern Recognition, 7464–7475. New York: IEEE.
Wang, W., C. Su, and D. Fu. 2022. “Automatic detection of defects in concrete structures based on deep learning.” Structures 43 (Dec): 192–199. https://doi.org/10.1016/j.istruc.2022.06.042.
Xie, J., Z. Guo, T. Wang, and J. Yang. 2023. “A diagnostic framework with a novel simulation data augmentation method for rail damages based on transfer learning.” Struct. Health Monit.-Int. J. 22 (5): 3437–3450. https://doi.org/10.1177/14759217221149129.
Yang, H., J. Liu, G. Mei, D. Yang, X. Deng, and C. Duan. 2023. “Research on real-time detection method of rail corrugation based on improved ShuffleNet V2.” Eng. Appl. Artif. Intell. 126 (Mar): 106825. https://doi.org/10.1016/j.engappai.2023.106825.
Yang, H., Y. Wang, J. Hu, J. He, Z. Yao, and Q. Bi. 2022. “Deep learning and machine vision-based inspection of rail surface defects.” IEEE Trans. Instrum. Meas. 71 (Jun): 1–14. https://doi.org/10.1109/TIM.2021.3138498.
Yu, J., X. Cheng, and Q. Li. 2022. “Surface defect detection of steel strips based on anchor-free network with channel attention and bidirectional feature fusion.” IEEE Trans. Instrum. Meas. 71 (Jun): 1–10. https://doi.org/10.1109/TIM.2021.3136183.
Zhang, B., S. Huang, L. Zhang, X. Li, X. Xu, and J. Lin. 2022. “Rail defect recognition based on waveform subtraction and rule base.” J. Perform. Constr. Facil. 36 (1): 04021101. https://doi.org/10.1061/(ASCE)CF.1943-5509.0001684.
Zhang, Z., S. Yu, S. Yang, and Z. Yu. 2021. “Rail-5k: A real-world dataset for rail surface defects detection.” Preprint, submitted June 28, 2021. http://arxiv.org/abs/2106.14366.
Zheng, Z., H. Qi, L. Zhuang, and Z. Zhang. 2021. “Automated rail surface crack analytics using deep data-driven models and transfer learning.” Sustainable Cities Soc. 70 (Jul): 102898. https://doi.org/10.1016/j.scs.2021.102898.
Information & Authors
Information
Published In
Copyright
© 2024 American Society of Civil Engineers.
History
Received: Feb 20, 2024
Accepted: May 22, 2024
Published online: Aug 13, 2024
Published in print: Oct 1, 2024
Discussion open until: Jan 13, 2025
ASCE Technical Topics:
- Algorithms
- Business management
- Continuum mechanics
- Defects and imperfections
- Dynamic loads
- Dynamics (solid mechanics)
- Engineering fundamentals
- Engineering mechanics
- Infrastructure
- Materials characterization
- Materials engineering
- Mathematics
- Practice and Profession
- Public administration
- Public health and safety
- Rail transportation
- Railroad tracks
- Railroad trains
- Safety
- Solid mechanics
- Structural dynamics
- Structural engineering
- Transportation engineering
- Vehicle loads
Authors
Metrics & Citations
Metrics
Citations
Download citation
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.