Real-Time Railway Track Defect Detection and Complaint Redressal Using YOLOv8
Main Article Content
Abstract
Railway infrastructure plays a critical role in modern transportation, where the timely identification of track defects is essential to ensure operational safety and efficiency. Traditional inspection techniques relying on manual surveys or basic image processing suffer from low scalability, delayed responses, and limited accuracy in detecting early-stage faults. This study proposes an intelligent, dual-module system for automated railway track defect detection and complaint redressal. The primary objective is to develop a robust real-time framework using YOLOv8 for accurate defect localization, coupled with an integrated backend to log and escalate issues for rapid maintenance response. The system utilizes the publicly available Railway Track Fault Detection (RTFD) dataset, comprising 5,000 annotated images across five defect classes. The YOLOv8 architecture is optimized with CSP-Darknet for feature extraction and PANet for multi-scale fusion, enabling high precision under diverse environmental conditions. Non-Maximum Suppression and IoU-based filtering are employed to eliminate false positives. A Python-Flask-based redressal module automatically logs defects with metadata into a PostgreSQL database and notifies maintenance teams. Experimental results demonstrate superior performance with mAP@0.5 of 93.4%, mAP@0.5:0.95 of 87.1%, precision of 92.8%, and real-time inference speed of 32 FPS. Compared to YOLOv5, YOLOv7, and CNN-based models, the proposed system delivers enhanced detection accuracy and operational readiness. In conclusion, this work presents a deployable, scalable, and efficient solution for AI-driven railway monitoring, bridging the gap between fault detection and actionable redressal, and paving the way for smart railway infrastructure.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
IJCERT Policy:
The published work presented in this paper is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. This means that the content of this paper can be shared, copied, and redistributed in any medium or format, as long as the original author is properly attributed. Additionally, any derivative works based on this paper must also be licensed under the same terms. This licensing agreement allows for broad dissemination and use of the work while maintaining the author's rights and recognition.
By submitting this paper to IJCERT, the author(s) agree to these licensing terms and confirm that the work is original and does not infringe on any third-party copyright or intellectual property rights.
References
M. Wang, K. Li, X. Zhu, and Y. Zhao, “Detection of Surface Defects on Railway Tracks Based on Deep Learning,” IEEE Access, vol. 10, pp. 15327–15338, 2022.
H. Hayre, “Automatic Railroad Track Inspection,” IEEE Transactions on Vehicular Technology, vol. 14, no. 1, pp. 58–63, 2020.
A. Kumar and R. Singh, “Railway Track Crack Detection Using Convolutional Neural Networks,” Proc. Int. Conf. on Smart Infrastructure, 2021.
S. Nguyen et al., “Autonomous Fault Detection in Railway Tracks Using Deep Learning,” IEEE Trans. Intelligent Transportation Systems, vol. 21, no. 6, pp. 3023–3032, 2020.
J. Lee and H. Park, “YOLOv7 and YOLOv8 in Real-Time Railway Track Defect Detection,” Journal of Computer Vision & AI, vol. 12, pp. 88–96, 2023.
G. Devi Appari et al., “Defect Detection Using YOLOv8 for Medical and Industrial Applications,” J. Theoretical & Applied Info Tech, vol. 102, no. 21, 2023.
E. Martinez and J. Johnson, “AI Models in Railway Infrastructure Monitoring: A Survey,” Machine Intelligence Review, vol. 4, no. 3, pp. 14–29, 2022.
C.-C. Hsieh, T.-Y. Hsu, and W.-H. Huang, "An Online Rail Track Fastener Classification System Based on YOLO Models," Sensors, vol. 22, no. 24, p. 9970, 2022.
M. Wang, K. Li, X. Zhu, and Y. Zhao, "Detection of Surface Defects on Railway Tracks Based on Deep Learning," IEEE Access, vol. 10, pp. 15327–15338, 2022.
J. Lee and H. Park, "YOLOv7 and YOLOv8 for Railway Track Defect Detection: A Comparative Study," Journal of Computer Vision & AI, vol. 12, pp. 88–96, 2023.
S. Nguyen et al., "Autonomous Fault Detection of Railway Tracks Using Deep Learning," IEEE Transactions on Intelligent Transportation Systems, vol. 21, no. 6, pp. 3023–3032, 2020.
G. D. Appari et al., "Detection and Classification of Diabetic Retinopathy Using YOLOv8 Deep Learning," Journal of Theoretical and Applied Information Technology, vol. 102, no. 21, 2023.
M. Barthakur and K. K. Sarma, "Semantic Segmentation of Satellite Images Using Deep Learning," Remote Sensing Analytics, vol. 10, no. 2, pp. 100–112, 2019.
Y. Shao, L. Zhang, and J. Liu, "LSTM-CNN Based Hybrid Model for Railway Defect Sequence Prediction," Proceedings of ICIRT, 2021.
S. Yadav and R. Patel, "Image Enhancement using GANs for Track Fault Detection," Computer Vision Trends, vol. 5, pp. 44–52, 2022.
W. Chen and M. Gao, "Lightweight Real-Time Detection of Rail Defects Using MobileNet-YOLO," IEEE Access, vol. 9, pp. 122456–122468, 2021.
S. Iyer, D. Gandhi, D. Mishra, R. Trivedi, and T. Ansa, "Solve My Problem: A Smart Grievance Redressal System," in Proc. SmartTech, pp. 134–144, 2023.
S. Enus, "Railway Track Fault Detection," Kaggle, 2019. [Online]. Available: https://www.kaggle.com/datasets/salmaneunus/railway-track-fault-detection [Accessed: 04-Apr-2025].
S. Chappidi and A. Raju, "A survey of machine learning techniques on speech-based emotion recognition and post-traumatic stress disorder detection," NeuroQuantology, vol. 20, no. 14, pp. 69–79, Oct. 2022, doi: 10.4704/nq.2022.20.14.NQ88010.
S. Chappidi and A. Raju, "Enhanced speech emotion recognition using the cognitive emotion fusion network for PTSD detection with a novel hybrid approach," Journal of Electrical Systems, doi: https://doi.org/10.52783/jes.644.
S. Chappidi and A. Raju, "Advancements in speech-based emotion recognition and PTSD detection through machine and deep learning techniques: A comprehensive survey," SSRG International Journal of Electronics and Communication Engineering, vol. 11, no. 5, 2023, doi: 10.14445/23488549/IJECE-V11I5P121.
S. Chappidi and A. Raju, "Speech-based emotion recognition by using a faster region-based convolutional neural network," Multimedia Tools and Applications, Springer, 2024, doi: https://doi.org/10.1007/s11042-024-19004-2.
M. Siahkouhi, J. Wang, X. Han, P. Aela, Y.-Q. Ni, and G. Jing, “Railway Ballast Track Hanging Sleeper Defect Detection Using a Smart Cnt Self-Sensing Concrete Railway Sleeper,” 2023, doi: 10.2139/ssrn.4414996.
M. S. Bellara, I. Chiha, and T. Ladhari, “Real-Time Face Detection and Recognition by the NAO Robot Using YOLO5 Algorithm,” 2024 IEEE International Conference on Artificial Intelligence & Green Energy (ICAIGE), pp. 1–6, Oct. 2024, doi: 10.1109/icaige62696.2024.10776694.
J. Sresakoolchai and S. Kaewunruen, “Railway defect detection based on track geometry using supervised and unsupervised machine learning,” Structural Health Monitoring, vol. 21, no. 4, pp. 1757–1767, Jan. 2022, doi: 10.1177/14759217211044492.