Accurate license plate localization

Abstract

Vehicle license plate identification system is an image-processing technology used to identify vehicles by their license plates. This technology is used in various security and traffic applications. This data is also used for enforcement, data collection, and can be used to keep a time record on the entry or exit of vehicles for automatic payment calculations. The significant advantage of this system is that it can keep an image record of the vehicle which is useful in order to fight crime and fraudulance. The key step in a vehicle license plate identification system, is plate region localization. This paper presents a new method to detect the location of license plate. The proposed algorithm consists of two main modules: license plate region's rough detection and accurate localization of region of interest(ROI). ROI is roughly estimated by vertical gradients. Rough ROI is utilized to obtain accurate ROI with the advent of DWT and morphological operations. The proposed method applies vertical subband feature of 2D discrete wavelet transform (DWT) to significantly highlight the vertical edges of license plates and suppress the surrounding background noise. Then, license plates can be extracted and located by the orthogonal projection histogram analysis with morphological operation. The proposed method can extract license plate of various vehicles both front and back. The experimental results show that the proposed method can achieve good license plate localization results with short runtime and high accurate detection rate.