An Improved Target Network Model for Rail Surface Defect Detection
Source abstract
Abstract
Rail surface defects typically serve as early indicators of railway malfunctions, which may compromise the quality and corrosion resistance of rails, thereby endangering the safe operation of trains. The timely detection of defects is essential to ensure the safe operation of railways. To improve the classification accuracy of rail surface defect detection, this paper proposes a rail surface defects detection algorithm based on MobileNet-YOLOv7. By integrating lightweight deep learning algorithms into the engineering application of rail surface defect detection, a MobileNetV3 lightweight network is used as the backbone network for YOLOv7 to enhance both speed and accuracy in complex defect extraction. Subsequently, the efficient intersection over union (EIOU) loss function is utilized as the positional loss function to bolster system resilience. Finally, the k-means++ clustering algorithm is applied to obtain new anchor boxes. The experimental results demonstrate the effectiveness of the proposed method, achieving superior detection accuracy compared with traditional algorithms.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Industrial Vision Systems and Defect Detection · Railway Engineering and Dynamics · Infrastructure Maintenance and Monitoring
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Beijing Engineering Research Center of Urban Transport Operation Guarantee, Beijing University of Technology, Beijing 100124, China
Beijing Key Laboratory of Traffic Engineering, Beijing University of Technology, Beijing 100124, China
School of Traffic and Transportation, Beijing Jiaotong University, Beijing 100044, China
State Key Laboratory of Advanced Rail Autonomous Operation, Beijing Jiaotong University, Beijing 100044, China
Funding
National Key Research and Development Pro-gram of China
References (39)
Show source-provided references
1. Automatic visual detection system of railway surface defects with curvature filter and improved Gaussian mixture model10.1109/TIM.2018.2803830 ↗
2. Intelligent recognition of defects in high-speed railway slab track with limited dataset10.1111/mice.13109 ↗
3. Automated rail surface crack analytics using deep data-driven models and transfer learning10.1016/j.scs.2021.102898 ↗
4. Faghih-Roohi, S., Hajizadeh, S., Nunez, A., Babuska, R., and De Schutter, B. (2016, January 24–29). Deep convolutional neural networks for detection of rail surface defects. Proceedings of the 2016 International Joint Conference on Neural Networks (IJCNN), Vancouver, BC, Canada.10.1109/IJCNN.2016.7727522 ↗
5. A visual detection system for rail surface defects10.1109/TSMCC.2012.2198814 ↗
6. Intelligent machine vision based railway infrastructure inspection and monitoring using UAV
7. Algorithms and instrument for rapid detection of rail surface defects and vertical short-wave irregularities based on fog and odometer10.1109/ACCESS.2019.2903488 ↗
8. Automatic detection of squats in railway infrastructure10.1109/TITS.2014.2307955 ↗
9. Characterization of defects on rail surface using eddy current technique10.1007/s12206-019-0816-x ↗
10. Railway track fastener defect detection based on image processing and deep learning techniques: A comparative study10.1016/j.engappai.2019.01.008 ↗
11. Ling, C., Guo, J., Gao, X., Wang, Z., and Li, J. (2016, January 22–24). Research on rail defect detection system based on FPGA. Proceedings of the 2016 IEEE Far East NDT New Technology & Application Forum (FENDT), Nanchang, China.10.1109/FENDT.2016.7992023 ↗
12. High-speed defect detection in rails by noncontact guided ultrasonic testing10.1177/0361198105191600110 ↗
13. Rail Crack Detection Using Optimal Local Mean Decomposition and Cepstral Information Coefficient Based on Electromagnetic Acoustic Emission Technology10.1109/TIM.2024.3375420 ↗
14. Dike, H.U., Zhou, Y., Deveerasetty, K.K., and Wu, Q. (2018, January 25–27). Unsupervised learning based on artificial neural network: A review. Proceedings of the 2018 IEEE International Conference on Cyborg and Bionic Systems (CBS), Shenzhen, China.10.1109/CBS.2018.8612259 ↗
15. MSRConvNet: Classification of railway track defects using multi-scale residual convolutional neural network10.1016/j.engappai.2023.105965 ↗
16. CUFuse: Camera and ultrasound data fusion for rail defect detection10.1109/TITS.2022.3189677 ↗
17. Sahebdivani, S., Arefi, H., and Maboudi, M. (2020). Rail track detection and projection-based 3D modeling from UAV point cloud. Sensors, 20.10.3390/s20185220 ↗
18. Automatic pavement texture recognition using lightweight few-shot learning10.1098/rsta.2022.0166 ↗
19. Real-time tunnel crack analysis system via deep learning10.1109/ACCESS.2019.2916330 ↗
20. Advancing Wildlife Protection: Mask R-CNN for Rail Track Identification and Unwanted Object Detection10.1109/ACCESS.2023.3313253 ↗
21. Task-efficiency oriented V2X communications: Digital twin meets mobile edge computing10.1109/MWC.012.2200465 ↗
22. Hybrid Transformer for Anomaly Detection on Railway HVAC Systems Through Feature Ensemble of Spatial–Temporal with Multi-channel GADF Images10.1007/s42835-024-01844-5 ↗
23. Multi-task Transfer Learning Facilitated by Segmentation and Denoising for Anomaly Detection of Rail Fasteners10.1007/s42835-022-01347-1 ↗
24. Automatic fastener classification and defect detection in vision-based railway inspection systems10.1109/TIM.2013.2283741 ↗
25. SeMA-UNet: A Semi-Supervised Learning with Multimodal Approach of UNet for Effective Segmentation of Key Components in Railway Images10.1007/s42835-024-01867-y ↗
26. Autonomous railway traffic object detection using feature-enhanced single-shot detector10.1109/ACCESS.2020.3015251 ↗
27. Machine learning-based rail corrugation recognition: A metro vehicle response and noise perspective10.1098/rsta.2022.0171 ↗
28. Defect classification based on deep features for railway tracks in sustainable transportation10.1016/j.asoc.2021.107706 ↗
29. Railway fastener inspection by real-time machine vision10.1109/TSMC.2014.2388435 ↗
30. Li, H., Wang, F., Liu, J., Song, H., Hou, Z., and Dai, P. (2022). Ensemble model for rail surface defects detection. PLoS ONE, 17.10.1371/journal.pone.0268518 ↗
31. Luo, H., Cai, L., and Li, C. (2023). Rail surface defect detection based on an improved YOLOv5s. Appl. Sci., 13.10.3390/app13127330 ↗
32. RAG-PaDiM: Residual attention guided PaDiM for defects segmentation in railway tracks10.1007/s42835-022-01346-2 ↗
33. Min, Y., Wang, Z., Liu, Y., and Wang, Z. (2023). FS-RSDD: Few-shot rail surface defect detection with prototype learning. Sensors, 23.10.3390/s23187894 ↗
34. A Detection System for Rail Defects Based on Machine Vision10.1088/1742-6596/1748/2/022012 ↗
35. Redmon, J., Divvala, S., Girshick, R., and Farhadi, A. (2016, January 27–30). You only look once: Unified, real-time object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA.10.1109/CVPR.2016.91 ↗
36. Wang, C., Bochkovskiy, A., and Liao, H. (2022). YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. arXiv.10.1109/CVPR52729.2023.00721 ↗
37. Howard, A.G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., and Adam, H. (2017). Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv.
38. Arthur, D., and Vassilvitskii, S. (2006). k-means++: The Advantages of Careful Seeding, Stanford.
39. Zoph, B., Cubuk, E.D., Ghiasi, G., Lin, T.-Y., Shlens, J., and Le, Q.V. (2020). Learning data augmentation strategies for object detection. Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, 23–28 August 2020, Springer International Publishing. Proceedings, Part XXVII 16.10.1007/978-3-030-58583-9_34 ↗