A Robust Model for Traffic Signs Recognition Based on Support Vector Machines

Road and traffic sign recognition has been of great interest for many years. This paper presents an approach to recognize Swedish road and traffic signs by using support vector machines. We focus on recognizing seven categories of traffic sign shapes and five categories of speed limit signs. Two kinds of features, binary image and Zernike moments, are used for representing the data to the SVM for training and test. We compare and analyze the performances of the SVM recognition model using different feature representations and different kernels and SVM types through recognizing 350 traffic sign shapes and 250 speed limit signs. Experiments have shown excellent results, which have achieved 100% accuracy on sign shapes classification and 99% accuracy on speed limit signs classification.