International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 11 | Nov 2025
p-ISSN: 2395-0072
www.irjet.net
Intelligent Traffic Signal Optimization Via Real -Time Density Estimation And Vehicle Counting With Canny Edge Detection And Yolov8 Prof.Shilpa Joshi 1, Basavaraj Malipatil2 1Professor, Master of Computer Application, VTU’s CPGS, Kalaburagi, Karnataka, India 2Student, Master of Computer Application, VTU’s CPGS, Kalaburagi, Karnataka, India
---------------------------------------------------------------------***--------------------------------------------------------------------reduce congestion and increase the efficiency of traffic Abstract - This article presents an intelligent traffic management.
signal optimization system that aims to improve urban traffic flow through automated perception and real-time decision-making. The system collects video data that includes vehicle detection, tracking, and density estimation, based on a model of deep learning and algorithms for tracking such as YOLOv8. These perception outputs capture traffic in real-time for each approach and are reliant upon a rigorous methodology for data fusion, accounting for varying light conditions and levels of congestion. The data is processed in a manner aligned with the max pressure, where green times are re-allocated based on real-time data collection and the software recognizes and maintains the optimal signal phases to mitigate wait time, congestion, and queue buildup. The system can communicate with existing controllers, and integrates a web-based dashboard for monitoring and supplementary manual controls. Periodic experimental trials illustrate significant improvements with respect to traffic responsiveness, levels of signal delay, and operational effectiveness as compared to fixed-time traffic signal control. This study provides a foundation for additional experimentation with integrating perception based on artificial intelligence with signal optimization based on adaptive planning for intelligent transportation systems.
2.PROBLEM STATEMENT Conventional traffic light systems utilize fixed timing cycles and are unable to adapt to real-time changes to vehicular flow. Because of this, intersections frequently suffer from high queue lengths, excessive delays and poor traffic flow - particularly occurring during peak hours. Productivity decreases while delays continue to grow without real-time traffic monitoring and adaptive control. Thus, there is a need for an intelligent system capable of estimating traffic density utilising modern computer vision technologies and optimally adjusting signal timings in real-time to enhance traffic flow efficiency and minimize delay. 3.OBJECTIVES The main goal of this initiative is to create a smart traffic signal control system which will self-adjust according to traffic conditions in real-time. The system will be based on computer vision methods to reliably detect and follow the movement of vehicles, track traffic density, and identify flow patterns at intersection locations. By combining these inputs, we will aim to optimize signal timings dynamically based on a suitable algorithm which will reduce congestion, waiting times, and improve overall traffic flows. Another goal of the initiative would be to present traffic information in an easily understandable dashboard that allows for monitoring and management of the system, while allowing for a straightforward integration into existing traffic control environments.
1.INTRODUCTION Traffic congestion is an expanding challenge in cities, often resulting in extensive delays and poor vehicle movement. Fixed-time traffic signals, the default for controlling car flow, cannot respond to fluctuating conditions making the traffic signal ineffective at peak hours or unexpected surges. Emerging technologies, specifically in computer vision and machine learning, make it possible to remotely and in real-time analyze traffic flow without on-site inspection using video based detection, tracking, and density estimation. The intelligent signal control system proposed here employs intelligent signal control techniques using computer vision and machine learning to develop simulation models that can assess car flow and dynamically apply signal timing using a max-pressure algorithm. The ultimate goal is to reduce wait times,
© 2025, IRJET
|
Impact Factor value: 8.315
4.RESEARCH METHODOLOGY The research method accentuates the creation of an intelligent traffic signal system through an organized and data-oriented procedure. The methodology consists of live video feeds being collected from traffic cameras and preparing the dataset for analysis. The vehicle detection and classification is then achieved using computer vision methodologies such as YOLOv8.. The detection task is followed by tracking algorithms to gauge vehicle
|
ISO 9001:2008 Certified Journal
|
Page 736