Published October 1, 2025 | Version v1.0
Preprint Open

Deep Learning Based Vehicle Tracking and Speed Estimation System in Restricted Traffic Zone

Authors/Creators

  • 1. Department of Computer Science and Engineering, C Abdul Hakeem College of Engineering and Technology, India

Contributors

  • 1. Assistant Professor, Department of Computer Science and Engineering, C Abdul Hakeem College of Engineering and Technology, India
  • 2. Professor & Head, Department of Computer Science and Engineering, C Abdul Hakeem College of Engineering and Technology, India

Description

This research introduces an artificial intelligence-driven framework designed to automate vehicle detection, tracking, and speed estimation in restricted traffic zones. The system leverages YOLOv8 for high-precision object detection, ByteTrack for robust multi-object tracking, and OpenCV for video processing within a Flask-based web application. A perspective transformation module maps pixel coordinates to real-world metrics, enabling accurate velocity and distance calculations critical for zones such as schools and hospitals.

Experimental results demonstrate a detection precision of 0.94, a tracking success rate of 0.91, and a speed estimation mean absolute error (MAE) of 2.5 km/h under varying traffic and environmental conditions. By integrating open-source technologies, the proposed system is scalable, cost-effective, and adaptable, contributing significantly to intelligent transportation systems and urban road safety.

Files

REAL-TIME DEEP LEARNING SYSTEM FOR SPEED & DISTANCE.pdf

Files (724.1 kB)

Additional details

Related works

Cites
Preprint: 10.48550/arXiv.2004.10934 (DOI)

Dates

Issued
2025-10-01
AI-driven system for vehicle detection, tracking, and speed estimation using YOLOv8, ByteTrack, and OpenCV, achieving high accuracy and scalability for intelligent traffic monitoring and urban safety.

Software

Repository URL
https://github.com/Salman-id85/AI-Powered-Vehicle-Tracking-and-Speed-Estimation
Programming language
Python
Development Status
Active

References

  • Bochkovskiy, A., Wang, C.-Y., & Liao, H.-Y. M. (2020). YOLOv4: Optimal Speed and Accuracy of Object Detection. arXiv:2004.10934. https://doi.org/10.48550/arXiv.2004.10934 Zhang, S., Liu, C., & Sun, F. (2021). Vehicle Detection and Speed Estimation using Faster R-CNN in Urban Traffic Surveillance. IEEE Transactions on Intelligent Transportation Systems, 22(3), 1352–1363. https://doi.org/10.1109/TITS.2020.2974321 Zhou, Y., Wang, D., & Krystian, M. (2021). ByteTrack: Multi-Object Tracking by Associating Every Detection Box. arXiv:2110.06864. https://doi.org/10.48550/arXiv.2110.06864 Hasan, M., Kabir, M. H., & Islam, S. M. S. (2022). Vision-Based Vehicle Detection and Speed Estimation Using YOLOv5 and DeepSORT. Proc. Int. Conf. Electrical, Computer, Communication Engineering (ECCE). Redmon, J., & Farhadi, A. (2018). YOLOv3: An Incremental Improvement. arXiv:1804.02767. https://doi.org/10.48550/arXiv.1804.02767