An Efficient Real-Time Traffic Sign Recognition System for Intelligent Vehicles with Smart Phones

摘要

In recent years, intelligent vehicles and smart phones have become more and more popular. The traffic sign recognition system is one kind of driving assistance system (DAS) which is used to automatically inform the driver the traffic sign information by a head up display (HUD), monitor, or speaker device. Hence this system is helpful to reduce driver distraction to increase driver safety. In this paper, an efficient real-time traffic sign recognition scheme is proposed for intelligent vehicles. The proposed scheme can integrate in-vehicle computing devices and smart phones to construe an in-vehicle traffic sign recognition system. This scheme contains four major stages: video frame capturing and transmitting, image preprocess, traffic sign detection, and character/icon extraction and recognition. The smart phone first captures videos and then extracts video frames in certain frame-rate. These extracted frames can be transmitted to an in-vehicle computing device by a wireless network (Bluetooth, WiMAX, Wi-Fi etc.) The preprocess employs some image processing to improve and transform the video frames to keep a stable quality for following detection and recognition schemes. At following stages, this paper presents some efficient and accurate traffic sign detection and recognition schemes which contain color selection, shape recognition, character/icon extraction and recognition. The experimental results shows that the proposed scheme only averagely spends 0.085 seconds for a video frame and its average accuracy can achieve about 98%. Hence, the proposed scheme can keep lower computing complexity, however it still can obtain a well accuracy.