Skip to main content

Advertisement

Springer Nature Link
Log in

Train track fastener defect detection algorithm based on MGSF-YOLO

  • Published:
The Journal of Supercomputing Aims and scope Submit manuscript

Abstract

The detection of train track fasteners is an important task to ensure the safe, reliable, and long-term stable operation of the railway system. Under the high stress and dynamic loads during train operation, track fasteners play a crucial role in maintaining the alignment and structural integrity of the tracks. However, traditional manual inspection methods are time-consuming, labor-intensive, and error-prone, while existing defect-detection algorithms often have the problem of low accuracy. To address these challenges, an improved YOLOv8n algorithm named MGSF-YOLO for high-precision detection of train track fastener defects is proposed. Firstly, an attention mechanism MSAM that combines MSCA and CBAM is proposed. MSAM adaptively adjusts the feature dimensions of different channels and the attention intensity at spatial positions through convolution kernels of various scales. This allows the model to focus on key features and regions and simultaneously enhances information extraction by exploiting the correlations between channels, thereby improving the model’s detection performance. Secondly, GSConv is used to replace Conv. This not only reduces the number of model parameters and computational cost but also improves the model’s detection accuracy. Then, an SPPFPool module that integrates average pooling and max pooling is proposed. This module aggregates the edge, background, and contextual information of the entire image, alleviates the impact of image scale changes, enhances the model’s adaptability to inputs of different sizes, effectively reduces false positives and false negatives, and improves the detection accuracy. Finally, the FocalerIoU loss function is introduced to replace the original CIoU loss function, which improves the regression accuracy and accelerates the model’s convergence speed. Experimental results show that compared with YOLOv8n, MGSF-YOLO has achieved performance improvements. The number of model parameters has decreased by 12.2%, and the computational amount has been reduced by 6.1%. Without sacrificing the detection speed, the mean average precision (mAP) has also increased by 2.6%. On the coco128 dataset, the publicly available NEU-DET dataset, and the real-world fastener dataset collected from the Roboflow website, the mAP of MGSF-YOLO has increased by 1.6%, 2.4%, and 3.6%, respectively, compared to YOLOv8n. Compared with other models, MGSF-YOLO reaches the highest mAP value, fully demonstrating its superior generalization ability and detection accuracy. It provides a reliable solution for the detection of railway fastener defects.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13
Fig. 14
Fig. 15
Fig. 16
Fig. 17
Fig. 18
Fig. 19
Fig. 20
Fig. 21
Fig. 22
Fig. 23
Fig. 24
Fig. 25
Fig. 26
Fig. 27
Fig. 28
Fig. 29
Fig. 30
Fig. 31
Fig. 32
Fig. 33
Fig. 34

Data Availability

The data used to support the findings of this study is available from the corresponding author upon request.

References

  1. Zhong H, Liu L, Wang J, Fu Q, Yi B (2022) A real-time railway fastener inspection method using the lightweight depth estimation network. Measurement 189:110613

    Article  MATH  Google Scholar 

  2. Meng Y, Yang J, Liu S, Xu W, Chen G, Niu Z, Wang M, Deng T, Qin Y, Han M et al (2022) Nano-fiber based self-powered flexible vibration sensor for rail fasteners tightness safety detection. Nano Energy 102:107667

    Article  Google Scholar 

  3. Hu Q, Fang F, Yuan R-J (2023) Fastener partial looseness identification of the ballastless track by a time-domain bayesian approach. Adv Struct Eng 26(10):1835–1846

    Article  MATH  Google Scholar 

  4. Li S, Jin L, Jiang J, Wang H, Nan Q, Sun L (2022) Looseness identification of track fasteners based on ultra-weak FBG sensing technology and convolutional autoencoder network. Sensors 22(15):5653

    Article  MATH  Google Scholar 

  5. Liu J, Huang Y, Zou Q, Tian M, Wang S, Zhao X, Dai P, Ren S (2019) Learning visual similarity for inspecting defective railway fasteners. IEEE Sens J 19(16):6844–6857

    Article  MATH  Google Scholar 

  6. Qiang W, Bailin L, Yun H, Hong F (2018) An improved LBP feature for rail fastener identification. J Southwest Jiaotong Univ 53(5):893–899

    MATH  Google Scholar 

  7. Bai T, Yang J, Xu G, Yao D (2021) An optimized railway fastener detection method based on modified faster R-CNN. Measurement 182:109742

    Article  MATH  Google Scholar 

  8. Ge X, Qin Y, Cao Z, Gao Y, Lian L, Bai J, Yu H (2023) A fine-grained method for detecting defects of track fasteners using rgb-d image. In: International Conference on Electrical and Information Technologies for Rail Transportation, Springer, pp 37–44

  9. Aydin I, Sevi M, Salur MU, Akin E (2022) Defect classification of railway fasteners using image preprocessing and alightweight convolutional neural network. Turk J Electr Eng Comput Sci 30(3):891–907

    Article  Google Scholar 

  10. Su S, Du S, Wei X, Lu X (2022) RFS-Net: railway track fastener segmentation network with shape guidance. IEEE Trans Circuits Syst Video Technol 33(3):1398–1412

    Article  MATH  Google Scholar 

  11. Sohan M, Sai Ram T, Reddy R, Venkata C (2024) A review on yolov8 and its advancements. In: International Conference on Data Intelligence and Cognitive Informatics, Springer, pp 529–545

  12. Wang L, Zang Q, Zhang K, Wu L (2024) A rail fastener defect detection algorithm based on improved yolov5. In: Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, 09544097241234380

  13. Hou Q, Zhou D, Feng J (2021) Coordinate attention for efficient mobile network design. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp 13713–13722

  14. Zhao Y, Ju Z, Sun T, Dong F, Li J, Yang R, Fu Q, Lian C, Shan P (2023) Tgc-yolov5: an enhanced yolov5 drone detection model based on transformer, gam & ca attention mechanism. Drones 7(7):446

    Article  Google Scholar 

  15. Yang L, Zhang R-Y, Li L, Xie X (2021) Simam: a simple, parameter-free attention module for convolutional neural networks. In: International Conference on Machine Learning, PMLR, pp 11863–11874

  16. Wang W, Dai J, Chen Z, Huang Z, Li Z, Zhu X, Hu X, Lu T, Lu L, Li H, et al (2023) Internimage: exploring large-scale vision foundation models with deformable convolutions. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp 14408–14419

  17. Tang J, Lei J, Xu D, Ma F, Jia K, Zhang L (2021) Sa-convonet: sign-agnostic optimization of convolutional occupancy networks. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp 6504–6513

  18. Li J, Wen Y, He L (2023) Scconv: spatial and channel reconstruction convolution for feature redundancy. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp 6153–6162

  19. Gupta C, Gill NS, Gulia P, Chatterjee JM (2023) A novel finetuned yolov6 transfer learning model for real-time object detection. J Real-Time Image Proc 20(3):42

    Article  MATH  Google Scholar 

  20. Wang K, Wei Z (2022) Yolo v4 with hybrid dilated convolution attention module for object detection in the aerial dataset. Int J Remote Sens 43(4):1323–1344

    Article  MATH  Google Scholar 

  21. Priya DK, Ramkumar M, Menaka D (2023) Rail fastener fault detection based on enhanced yolov7 model. In: 2023 International Conference on Circuit Power and Computing Technologies (ICCPCT), IEEE, pp 1681–1689

  22. Wang C-Y, Yeh I-H, Mark Liao H-Y (2025) Yolov9: Learning what you want to learn using programmable gradient information. In: European Conference on Computer Vision, Springer, pp 1–21

  23. Zhang Y-F, Ren W, Zhang Z, Jia Z, Wang L, Tan T (2022) Focal and efficient IOU loss for accurate bounding box regression. Neurocomputing 506:146–157

    Article  MATH  Google Scholar 

  24. Zou H, Wang Z (2024) An enhanced object detection network for ship target detection in SAR images. J Supercomput 80:17377–17399

    Article  Google Scholar 

  25. Guo M-H, Lu C-Z, Hou Q, Liu Z, Cheng M-M, Hu S-M (2022) Segnext: rethinking convolutional attention design for semantic segmentation. Adv Neural Inf Process Syst 35:1140–1156

    MATH  Google Scholar 

  26. Li H, Li J, Wei H, Liu Z, Zhan Z, Ren Q (2024) Slim-neck by gsconv: a lightweight-design for real-time detector architectures. J Real-Time Image Proc 21(3):62

    Article  MATH  Google Scholar 

  27. Tian D, Yan X, Zhou D, Wang C, Zhang W (2024) Iv-yolo: a lightweight dual-branch object detection network. Sensors 24(19):6181

    Article  MATH  Google Scholar 

  28. Cheng Z, Gao L, Wang Y, Deng Z, Tao Y (2024) Ec-yolo: effectual detection model for steel strip surface defects based on yolo-v5. IEEE Access 12:62765–62778

  29. Fu J, Chen X, Lv Z (2022) Rail fastener status detection based on mobilenet-yolov4. Electronics 11(22):3677

    Article  Google Scholar 

  30. Li X, Wang Q, Yang X, Wang K, Zhang H (2023) Track fastener defect detection model based on improved yolov5s. Sensors 23(14):6457

    Article  MATH  Google Scholar 

  31. Li N, Wang M, Yang G, Li B, Yuan B, Xu S (2024) Dens-yolov6: a small object detection model for garbage detection on water surface. Multimed Tools Appl 83(18):55751–55771

    Article  MATH  Google Scholar 

  32. Wang C-Y, Bochkovskiy A, Liao H-YM (2023) Yolov7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp 7464–7475

  33. Liu J, Liu H, Chakraborty C, Yu K, Shao X, Ma Z (2021) Cascade learning embedded vision inspection of rail fastener by using a fault detection IoT vehicle. IEEE Internet Things J 10(4):3006–3017

    Article  Google Scholar 

  34. Wang C-Y, Liao H-YM et al (2024) Yolov1 to yolov10: the fastest and most accurate real-time object detection systems. APSIPA Trans Signal Inf Process 13(1):1–13

    Article  MATH  Google Scholar 

  35. Wang Z, Su Y, Kang F, Wang L, Lin Y, Wu Q, Li H, Cai Z (2025) Pc-yolo11s: a lightweight and effective feature extraction method for small target image detection. Sensors 25(2):348

    Article  MATH  Google Scholar 

Download references

Funding

This study was supported by the Key Laboratory Fund Project of National Defense Science and Technology (2022-JCJQ-L8-015-020), the Key Project of Scientific Research Project of Liaoning Provincial Department of Education (LJKZ0475), and the Innovation Support Program for High-level Talents of Dalian City (2022RJ03).

Author information

Author notes

  1. Ronghua Li and Henan Hu have been contributed equally to this work.

Authors and Affiliations

  1. School of Mechanical Engineering, Dalian Jiaotong University, No. 794, Yellow River Road, Dalian, 116028, Liaoning, China

    Siwei Ma & Henan Hu

  2. School of Automation and Electrical Engineering, Dalian Jiaotong University, No. 216 Xingfa Road, Lushunkou Economic Development Zone, Dalian, 116045, Liaoning, China

    Ronghua Li

Authors
  1. Siwei Ma
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Ronghua Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Henan Hu
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

All authors were involved in the conception and design of the algorithms, data collection and analysis was done by RHL, theoretical analysis and experiments were done by SWM and HNH, the first draft of the manuscript was written by SWM and HNH, and all authors have read and agreed to the published version of the manuscript.

Corresponding author

Correspondence to Ronghua Li.

Ethics declarations

Conflict of interest

The authors declare no competing interests.

Ethical approval and informed consent

This article does not contain studies with human participants or animals. Statement of informed consent is not applicable since the manuscript does not contain any patient data.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ma, S., Li, R. & Hu, H. Train track fastener defect detection algorithm based on MGSF-YOLO. J Supercomput 81, 494 (2025). https://doi.org/10.1007/s11227-025-07024-0

Download citation

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s11227-025-07024-0

Keywords