Real-Time Traffic Sign Recognition Using Integrated Camera Sensors and Yolov8 Algorithm

Abstract

Traffic signs are essential in maintaining smoothness and safety on the road. However, many drivers still violate them, causing various negative impacts. Traffic Sign Recognition (TSR) is a technology that detects and identifies various types of traffic signs by utilizing artificial intelligence in the domain of computer vision. TSR has been applied in various vehicle applications, such as the Advance Driver Assistance System (ADAS) and Autonomous Driving System (ADS). This system integrates camera sensors with the YOLOv8 algorithm, which has high accuracy and fast data processing. The data used were 2093 images and annotated through Roboflow. Then, data were trained through Google Collaboratory with a mAP evaluation result of 95.5%, showing that the system can detect objects accurately. The precision of the model in detecting objects was 93.5%, while its success rate was 93.3%. The testing results of the system can utilize images, videos, and cameras in real time.