Traffic Sign Recognition Using Deep Learning

Abstract

Traffic Sign Recognition is the process of automatically detecting and classifying traffic signs along the road, including speed limit signs, yield signs, merge signs etc. It can subsequently be used in ADAS (advanced driver-assistance systems), or for help in the navigation of self-driving cars. It is a technology that allows users to recognize traffic signs in real-time, typically in videos, or sometimes just in photos. It is a realistic task that is full of constraints, such as visual environment, physical damages, and partial occasions, etc. In order to deal with the constrains, convolutional neural networks (CNN) are accommodated to extract visual features of traffic signs and classify them into corresponding classes. Our final-year project automatically and in real time uses Deep Convolutional Neural Networks (CNNs) to achieve the task of automated traffic sign detection and classification. We initially used a benchmark (GTSRB) for the traffic-sign recognition. In order to determine which deep learning models are the most suitable one for the TSR (Traffic Sign Recognition), we chose three kinds of models to conduct deep learning computations: LENET, SSD MobileNet and YOLOv4. According to the scores of various metrics, we summarized the pros and cons of the picked models for the TSR task. We have used the Python programming language, OpenCV library and Linux OS. The NVIDIA Jetson Nano single-board computer is used for hardware deployment.