Published in: Journal of Computer Science and Engineering in Innovations and Research (ISSN No: 3049-1762 online)
Publication Date: May 15, 2025
This project introduces a sophisticated Traffic Monitoring System integrating YOLOv9 for precise vehicle detection and DeepSORT for continuous multi-object tracking, ensuring that vehicles remain uniquely identified across multiple frames. Traditional CCTV surveillance systems mainly function as passive monitoring tools and lack the intelligence to track suspicious vehicles in real-time. This constraint makes a vehicle that is supposed to be tracked go undetected once it is out of view.
The proposed system fills this critical gap as it tracks in real-time and monitors continuously so that no vehicle of interest is ever missed. The system applies perspective transformation, computes time-distance, and gets accurate speed on roads with varying geometries. This would allow for the detection of erratic driving behaviors like overspeeding or abrupt braking.
A real-time dashboard, developed using Streamlit, presents key insights into traffic through live video feeds, vehicle counts, speed distributions, etc. This urban environment-friendly system is scalable on multiple CCTV cameras, making city-wide or highway-wide tracking possible without breaking up the continuity of tracking.
This tool is not just a prerequisite in the sense of traffic flow analyses and congestion control but also plays a crucial role in the proactive management of law and order, making it highly essential for smart city infrastructure.
Traffic Monitoring System, Vehicle Detection, Multi-Object Tracking, Real-Time Surveillance, Smart City Infrastructure