A robust algorithm for speed limit sign detection

Alerting the driver of the speed limit is an important measure to keep the road safe. It is a challenging engineering problem to develop an automatic system which can detect speed limit signs under ever changing road conditions. In this paper, a robust algorithm for detecting speed limit sign is developed based on the traditional framework of pre-processing, feature extraction and detection. There are two major contributions of this paper. Based on studying the step response of the log ratio between the arithmetic mean and the geometric mean, we propose a new feature extraction scheme which is completely different from the gradient-based edge detection schemes. We also develop a sequential training algorithm for detecting the target using a simple feed-forward network. Other contributions of this paper include selecting possible targets using a median operator and ranking the possible targets based on the output from a feed forward network. Computer experiments have been carried out to verify the performance of the proposed algorithm which successfully detects speed limit signs in seven video sequences provided by the Co-operative Research Centre for Advanced Automotive Technology (AutoCRC), Australia. Preprint submitted to Elsevier June 28, 2007