A novel smart multi-license plate recognition method

Abstract

This paper presents a Smart License Plate Recognition System (SLPRS) which employs a cascade scheme guaranteeing high accuracy and reliability. The system consists of three phases. First, the proposed pseudo-morphological closing operation algorithm is applied to extract license plate candidates based on contour image and the so-called license plate transition rule. Second, improved adaptive template-matching algorithm is used in character segmentation to obtain precise segmented characters by solving an optimization model. Finally we adopt pattern match method for recognition based on proper designed character templates. Compared to the traditional morphology method, connected component analysis and projection method, our method is lower computational, more accurate and parametric intelligent, which reaches the basic requirements of SLPRS. Experiments show that the system can recognize license plates in various sizes of images with different scenes without changing the parameters in the system. © 2012 Springer-Verlag.