Home / Current Issue / Paper 1719091
CNN-Based Traffic Signal Detection for Low Visibility Scenarios
Subject area: Science,Engineering and Technology · Area of research: Machine Learning
DOI: 10.64388/IREV9I12-1719091
Abstract
Traffic signal detection is a crucial component of intelligent transportation systems (ITS), enabling safer and more efficient road usage. However, accurately identifying traffic signals under adverse environmental conditions such as fog, rain, and low-light environments remains a significant challenge. Traditional image processing techniques often fail to provide reliable results due to poor visibility and environmental noise. This project presents a Convolutional Neural Network (CNN)-based approach for robust traffic signal detection in low-visibility scenarios. The proposed system utilizes a deep learning model trained on a diverse dataset of traffic signal images collected under various weather conditions. Preprocessing techniques such as image resizing, normalization, and contrast enhancement are applied to improve feature extraction and model performance. The model is developed using Python with TensorFlow and scikit-learn, and evaluated based on performance metrics including accuracy, precision, and recall. The system is designed to classify traffic signals into different categories such as red, yellow, and green with high reliability. A simple graphical user interface (GUI) is also implemented using Tkinter to facilitate user interaction. The expected outcome of the project is an efficient and accurate traffic signal detection system capable of operating under challenging environmental conditions. This system can be further integrated into autonomous vehicles, driver assistance systems, and smart traffic management solutions to enhance road safety and automation.
References
[1] Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet Classification with Deep Convolutional Neural Networks,” Communications of the ACM, vol. 60, no. 6, pp. 84–90, 2017.
[2] K. Simonyan and A. Zisserman, “Very Deep Convolutional Networks for Large-Scale Image Recognition,” arXiv preprint arXiv:1409.1556, 2014.
[3] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real- Time Object Detection,” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
[4] P. Sharma, A. Gupta, and R. Singh, “Traffic Light Detection under Adverse Weather Conditions Using CNN,” IEEE Access, vol. 9, pp. 112430–112441, 2021.
[5] M. Hossain and A. Rahman, “Improved Traffic Light Recognition Using Convolutional Neural Networks under Low Visibility Conditions,” Scientific Reports, vol. 13, no. 5, pp. 8891– 8899, 2023.
[6] D. Chen and F. Luo, “Enhancing Road Safety Using CNN-Based Traffic Light Detection System,” Scientific Reports, vol. 14, no. 2, pp. 2567–2576, 2024.
[7] R. Jain and A. Sinha, “Detection of Traffic Lights Using Transfer Learning in Adverse Conditions,” IEEE Region 10 Conference (TENCON), 2022.
[8] X. Zhang, P. Sun, and J. Wang, “Multi-Weather Condition Robust Traffic Signal Detection Using CNNs,” IEEE Access, vol. 10, pp. 55341–55352, 2022.
[9] R. Patel and K. Shah, “Smart Traffic Management Using CNN -Based Signal Detection and SQLite Integration,” International Journal of Emerging Trends in Engineering Research, vol. 12, no. 8, pp. 145–152, 2024.
[10] S. Mehta, D. Patel, and A. Bhatt, “Foggy and Rainy Weather Image Classification Using CNN and Data Augmentation Techniques,” International Journal of Computer Applications, vol. 185, no. 2, pp. 10–18, 2023.
How to cite this paper
@article{1719091,
author = {Rohan Thanage, Onkar Shete, Utkarsh Tope, Dr. D. V. Gore},
title = {CNN-Based Traffic Signal Detection for Low Visibility Scenarios},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {12},
pages = {2695-2700},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1719091.pdf},
abstract = {Traffic signal detection is a crucial component of intelligent transportation systems (ITS), enabling safer and more efficient road usage. However, accurately identifying traffic signals under adverse environmental conditions such as fog, rain, and low-light environments remains a significant challenge. Traditional image processing techniques often fail to provide reliable results due to poor visibility and environmental noise. This project presents a Convolutional Neural Network (CNN)-based approach for robust traffic signal detection in low-visibility scenarios. The proposed system utilizes a deep learning model trained on a diverse dataset of traffic signal images collected under various weather conditions. Preprocessing techniques such as image resizing, normalization, and contrast enhancement are applied to improve feature extraction and model performance. The model is developed using Python with TensorFlow and scikit-learn, and evaluated based on performance metrics including accuracy, precision, and recall. The system is designed to classify traffic signals into different categories such as red, yellow, and green with high reliability. A simple graphical user interface (GUI) is also implemented using Tkinter to facilitate user interaction. The expected outcome of the project is an efficient and accurate traffic signal detection system capable of operating under challenging environmental conditions. This system can be further integrated into autonomous vehicles, driver assistance systems, and smart traffic management solutions to enhance road safety and automation.},
month = {June},
doi = {https://doi.org/10.64388/IREV9I12-1719091}
}