Smart Traffic Management System using Image Processing

Abstract

Delayed traffic and a rise in fuel contamination in cities happen due to fixed-time traffic signals that do not adapt to demand. As a result, a new traffic control system that combines Deep Learning and Machine Learning has been designed. Using predefined YOLOv8 Nano and EfficientDet-D0 models, the video from surveillance cameras is examined to find vehicles and determine their speed. YOLOv8 Nano was selected because it’s fast and accurate which lets us input its predictions into a Random Forest Regressor that controls the green light timing. Because this system focuses on busier places, it decreases wait times and helps signals run more efficiently. Because of this, cities use much less energy and generate less air pollution, contributing to smarter city designs.

Introduction

With increasing urbanization, city traffic congestion has become a major problem. Traditional traffic systems rely on fixed signal timings, which do not adapt to real-time conditions. This leads to unnecessary vehicle idling, fuel consumption, and pollution.

To address these issues, this study proposes an AI-based Smart Traffic Management System using deep learning and machine learning to optimize traffic flow in real time.

II. Background:

• Urban growth has drastically increased vehicle density.

• Fixed-timing traffic signals are inefficient and contribute to longer wait times, environmental damage, and resource waste.

• Adaptive systems are needed to respond to changing traffic conditions dynamically.

• Advancements in AI and image processing allow for more intelligent traffic control solutions.

III. Methodology:

The proposed system uses object detection and predictive modeling to dynamically adjust traffic lights based on real-time data.

Key Components:

1. YOLOv8 Nano:

   • Lightweight, fast object detection model.

   • Detects vehicles and people in real-time with bounding boxes and confidence scores.

   • Ideal for real-time deployment on low-power devices.

2. EfficientDet-D0:

   • Uses BiFPN and EfficientNet for multi-scale, high-accuracy detection.

   • Detects small or closely packed objects but is more computationally demanding.

3. Model Selection:

   • YOLOv8 Nano chosen for its speed, efficiency, and precision in real-time applications.

4. Random Forest Regressor:

   • Predicts optimal green light durations.

   • Uses features like vehicle count and congestion level for dynamic signal adjustments.

IV. Proposed System:

An autonomous traffic control system that:

• Processes live video from intersection cameras.

• Detects vehicles using YOLOv8 Nano.

• Predicts traffic congestion and vehicle speed.

• Adjusts signal timings using a Random Forest model to reduce queueing and wait times.

• Operates without human intervention, improving traffic efficiency, reducing emissions, and enhancing road safety.

V. Implementation:

Three main algorithms form the system:

1. YOLOv8 Nano:

• Detects and tracks vehicles (cars, buses, trucks, motorcycles).

• Calculates speed and congestion using vehicle movement across frames.

2. EfficientDet-D0:

• Used for complex scenes with small objects.

• Detects and tracks vehicles, but is slower and more resource-intensive.

3. Combined YOLOv8 + Random Forest:

• YOLOv8 detects vehicles and congestion.

• Random Forest predicts the best green signal time based on live traffic conditions.

• Overlays control decisions on video for visualization.

? Key Benefits:

• Real-time and adaptive signal control.

• Reduces fuel use, pollution, and idle time.

• Improves traffic flow, safety, and efficiency in urban areas.

• Scalable to future smart city infrastructure.

Conclusion

In this study, a smart traffic management system with real time vehicle detection implemented using YOLOv8, EfficientDet and traffic signal optimization utilizing a Random Forest Regressor is presented. Based on live traffic flow and speed the system adapts signal timings, resulting in decreased congestion, less waiting time, less emissions and more traffic efficiency. Future work involves embedding road sensors and IoT devices to improve precision, coordinating multiple intersections for continuous traffic flow, adjusting signal behavior according to weather conditions, giving emergency vehicles the first right of way and designing a decentralized IoT based architecture for smarter traffic control city wide.

References

[1] A. K., M. H., S. P., and P. D. R., “Automated Traffic Management Using Image Processing,” in IEEE International Conference on ‘New Frontiers in Communication, Automation, Management and Security (ICCAMS-2023)’, 2023 [2] G. Niharika, N. Sathanya, Rahena, T. Akshitha, and P. Madhuri, “Vehicles Congestion Controller and Penalty System,” in International Journal of Advances in Engineering and Management (IJAEM), 2023. [3] K. Senthilkumar, V. Ellappan, and A. A. R. Arun, “Traffic Analysis and Control Using Image Processing,” in IOP Conference Series: Materials Science and Engineering, 2017. [4] M. Gupta, H. Miglani, P. Deo, and A. Barhate, “Real-time Traffic Control and Monitoring,” in e-Prime - Advances in Electrical Engineering, Electronics, and Energy (Elsevier), 2023. [5] P. N. Chowdhury, T. C. Ray, and J. Uddin, “A Vehicle Detection Technique for Traffic Management using Image Processing,” in International Conference on Electrical, Computer and Communication Engineering (ECCE), 2017. [6] P. Roopa Ranjani, B. Sneha Sri, G. Srija Reddy, P. Sonali, and V. Vaishnavi, “Traffic Detection and Estimation Using Image Processing,” in International Journal of Creative Research Thoughts (IJCRT), 2023. [7] P. Akanksha, T. Divya Kumari, and R. Mamatha, “Traffic Density Detection using Video Tracking,” in International Journal of Analytical and Experimental Modal Analysis, 2020. [8] S. Agnes Joshy, M. Dhana Sakhi, C. Jenit Catherin, and T. Danamaliga, “Traffic Management System Using Machine Learning,” in International Journal of Novel Research and Development (IJNRD), 2024 [9] S. S. Tippannavar, S. Jain, R. Harshith, and S. S. Chandanagiri, “Image Processing Based Intelligent Traffic Control System Using OpenCV,” in International Journal of Creative Research Thoughts (IJCRT), 2022. [10] V. Bhardwaj, Y. Rasamsetti, and V. Valsan, “Image Processing Based Smart Traffic Control System for Smart City,” in 12th International Conference on Computing, Communication, and Networking Technologies (ICCCNT), IEEE, 2021.

Copyright

Copyright © 2025 Mrs. Mallepally Anusha, Addula Rishitha, Kolli Thanmai , Poojitha Sri Gayathri Pragallapati. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.