A Lightweight Model For Real-time Traffic Sign Recognition

Abstract

Traffic sign detection (TSR) is a hot topic in the field of computer vision with lots of applications such as autonomous vehicles, path planning, robot navigation etc. Especially, it also can be applied in advanced driver assistance system (ADAS) which can help drivers get more useful information on the road to make decision exactly. However, most of the developed systems cannot be used in real-time environment. Therefore, in this paper, a light-weight model has been proposed for real-time traffic sign recognition. Our system is embedded on a 1:7 RC vehicle and tested on our small driving environment. The experimental results demonstrate the effectiveness and robustness of the proposed model in many challenging situations.