Cheung C, Yan PF, and Tam WP, Smart Guardian: creating a secure railway through the discovery of suspicious objects, HKIE Transactions, Vol. 32, No. 1 (Regular Issue), Article THIE-2025-0013.R2, 2026, 10.33430/V32N1THIE-2025-0013
Many railway operations rely on staff vigilance and basic security measures to maintain safety in busy areas like concourses where passengers start their journeys. However, growing passenger numbers and complex railway environments make it harder to spot threats like unattended bags or prohibited items. Safety is paramount in railway operations and critical to public infrastructure. Therefore, an innovative autonomous system called the Smart Guardian, employing artificial intelligence and computer vision, was explored to automatically detect suspicious objects. The system relies on deep learning algorithms trained on railway-specific data to handle crowded and dynamic scenes. It works better than conventional methods because it can overcome challenges like moving foreground objects and busy backgrounds. The system analyses CCTV footage in real time and quickly spots prohibited items and unattended luggage. Once a potential threat is identified, the system immediately alerts station operators, enabling them to take appropriate and timely actions to address the situation, mitigate safety and security risks, and ensure the safety of passengers and staff. With its advanced detection capabilities, the system could become a crucial tool in safeguarding public spaces and transportation while preventing potential security incidents.
Keywords:
Artificial intelligence; computer vision; video analytic; railway; safety; surveillance
Reference List:
1. (2024). MTR. [online]. Available at: . [Accessed on 21 January 2025].
2. (2024). MTR. [online]. Available at: . [Accessed on 21 January 2025].
3. Zucchi, K. (2022). What countries spend on antiterrorism. [online]. Available at: [Accessed on 21
January 2025].
4. (2023). Executive Summary. In: Global Terrorism Index 2023. [online]. Available at: [Accessed on 21 January
2025].
5. Albawi, S., Mohammed, T. A., & Al-Zawi, S. (2017). Understanding of a convolutional neural network. In 2017 international conference on engineering and technology (ICET) (pp. 1-6). Ieee.
6. Schulzrinne, H., Rao, A., & Lanphier, R. (1998). RFC2326: Real time streaming protocol (RTSP).
7. Masse, M. (2011). REST API design rulebook: designing consistent RESTful web service interfaces." O'Reilly Media, Inc.".
8. Jiang, P., Ergu, D., Liu, F., Cai, Y., & Ma, B. (2022). A Review of Yolo algorithm developments. Procedia Computer Science, 199, 1066-1073.
9. (2023). Person and Luggage Object Detection Dataset (v2, coco classes - person, backpack, handbag, suitcase). In: Roboflow Universe. 1st ed. [online]. Available at: . [Accessed on 21 January 2025].
10. Xu, Y., Dong, J., Zhang, B., & Xu, D. (2016). Background modeling methods in video analysis: A review and comparative evaluation. CAAI Transactions on Intelligence Technology, 1(1), 43-60.
11. Stauffer, C., & Grimson, W. E. L. (1999). Adaptive background mixture models for real-time tracking. In Proceedings. 1999 IEEE computer society conference on computer vision and pattern recognition (Cat. No PR00149) (Vol. 2, pp. 246-252). IEEE.
12. (2025). Blender. [online]. Available at: . [Accessed on 21 January 2025].
13. Weiss, K., Khoshgoftaar, T. M., & Wang, D. (2016). A survey of transfer learning. Journal of Big data, 3(1), 1-40.
14. Henderson, P., & Ferrari, V. (2017). End-to-end training of object class detectors for mean average precision. In Computer Vision–ACCV 2016: 13th Asian Conference on Computer Vision, Taipei, Taiwan, November 20-24, 2016, Revised Selected
Papers, Part V 13 (pp. 198-213). Springer International Publishing.
15. Yacouby, R., & Axman, D. (2020). Probabilistic extension of precision, recall, and f1 score for more thorough evaluation of classification models. In Proceedings of the first workshop on evaluation and comparison of NLP systems (pp. 79-91).
16. Villán, A.F. 2019. Mastering OpenCV 4 with Python: A practical guide covering topics from image processing, augmented reality to deep learning with OpenCV 4 and Python 3.7. Packt Publishing.
17. (2025). PyTorch. [online]. Available at: . [Accessed on 21 January 2025].
