Speed limit traffic sign detection and recognition

Since carelessness about the maximum speed allowed in roads has caused many fatal accidents, an intelligent system to detect and inform it automatically, seems to be very useful. In this paper, a system is developed to detect and recognize the speed limit traffic signs, which takes digital images from driving scenes and then the pictures are being filtered and searched to find the circular speed limit sign. The new method proposed, that is named "hierarchical Hough transform" can optimize the solution of accuracy/complexity problem. The powerful tool of neural network is then applied to recognize the pattern (number) which has been well separated by the previous method. Some results from natural scenes are shown during the process presentation