Car license plate recognition based on the combination of principal components analysis and radial basis function networks

This paper proposes a car license plate recognition system. First, an approach based on histogram analysis method together with Hough transform is presented for segmentation. It is robust to car plate inclination within 15 degree. Subsequently, in the recognition phase, we use principle components analysis to extract the feature of input images and then feed to neural networks with radial basis function networks is used for symbols recognition. Experimental results show that the PCA method proposed reduces the dimension of image effectively and the recognition performance using RBF networks with 98% is better than back-propagation (BP) with 93%.