ABSTRACT
Traffic sign detection is a critical component in the development of intelligent transportation systems and autonomous driving technologies. This project presents a robust approach to detecting traffic signs using the YOLO (You Only Look Once) object detection algorithm. By leveraging a custom-trained YOLOv5 model on a dataset of traffic sign images, the system is capable of identifying and localizing multiple classes of signs in real-time with high accuracy. The model was trained and evaluated using annotated images and demonstrates effective detection performance under varying lighting and environmental conditions. The application supports real-time testing via webcam or video input, showcasing its potential for integration into advanced driver-assistance systems (ADAS) or smart vehicular platforms. The project highlights the advantages of using YOLO for real-time object detection tasks and contributes to safer and more intelligent road navigation systems.
Keywords:
Traffic Sign Detection, YOLOv5, Real-time Detection, Deep Learning, Convolutional Neural Networks (CNN)


OBJECTIVES
1. To collect and preprocess a dataset consisting of labeled images of various traffic sign classes under different lighting and environmental conditions.
2. To train a custom YOLOv5 model for the task of multi-class traffic sign detection, optimizing for both accuracy and inference speed.
3. To evaluate the model's performance using standard object detection metrics such as precision, recall, and mean Average Precision (mAP).
4. To implement real-time detection capabilities using video input or live webcam feed to demonstrate the practical applicability of the trained model.
5. To analyze and interpret the results with a focus on model performance, limitations, and potential improvements for future work.
6. To contribute to the development of intelligent transportation systems by providing a scalable and deployable solution for traffic sign recognition.