Research Of The Vehicle License Plate Location Methods Based On Features

The license plate recognition system (LPR) is the main part of modern Intelligent Transportation System (ITS).License plate location is a prerequisite for license plate recognition. Though LPR is a mature technology in fixed imaging environment, the existing LPR technologies do not meet the special mission requirements in complex imaging environment, especially in mobile conditions. This paper proposed an algorithm of license plate location in complicated imaging environment, which is fit for multi-perspective, and various illuminating conditions.This method uses various features of license plate, such as color, geometric, structure and gray correlations. from coarse to fine to narrow the searching range. It can locate the license plate in complex imaging environment.The method contains four parts:1. Extraction the region of license plate based on license plate's color feature.2. Locate the license plate based on region object regional-level segmentation.3. Locate the license plate based on region object features clustering.4. Confirming the region of license plate based on geometric correction, horizontal and vertical projection.Experiment results show that this method has a strong adaptability in many kinds of imaging environment. This paper describes the basic principle and implementation of this method. Finally, the experimental results are analyzed.