Road Traffic Sign Detection

Road traffic signs provide important information for guiding, warning, or regulating the behaviors drivers in order to make driving safer and easier. As one of the major tasks of driving assistance systems, road traffic sign detection and recognition has gain more and more popularity since recent years. This topic is a great vision task since there are a number of important issues that need to be addressed.This thesis begins with an instruction to the development of road traffic sign detection and recognition technology and follows a brief description of several primary methods for traffic sign detection. We begin color segment and carry up a research into the use of SVF (Simple Vector Filter) in segmentation of the color of traffic sign. We made a comparison between the experiment result of SVF and color detection based on channel H in HSV space, and improves the algorithm by excluding the green part.Different shape analysis is done in region with distinct color features. Rectangular signs and triangle signs are detected using corner extraction and geometric feature analysis while circular signs are searched by a new kind of codification based on the symmetry nature of the sign and the existence of its neighbors. This method locates and gets the size of the circle by searching symmetric points in the image. Experiments show good results of detecting road traffic signs under complex backgrounds and different weather conditions. Besides, this method has competitive accuracy in sign detection in complex scene images and favorable speed for a real-time system.We also carry out some experiments in the recognition of forbidden signs using the feature based-method. The detection is a basic step for further recognition.