Traffic Sign Recognition

Smart vehicles with capabilities of autonomous driving are a big revolution in automobile industry. The vehicles can sense their environment and react based on it. It replaces the manual driver. Recognition of traffic sign is an important enabler for autonomous driving. Camera installed in the vehicle captures the traffic sign on the road and they must be recognized accurately for triggering the suitable action. In this thesis both image processing and Euclidean distance matching are used to pre-process and classify the traffic signs and thresholding and thinning are applied on image for feature extraction. In this work, a simple, efficient traffic sign recognition system with low computational time and to achieve good accuracy is proposed. Time to classify the traffic sign is achieved in milliseconds and accuracy is maintained using the proposed system. Keywords: Autonomous Driving, Image processing, Thresholding, Thinning, Euclidean distance matching