Research And Realization Of Licencse Plate Recgniton Algorith

In recent years, with the development of the social economy and the improvement of thepeople's living standards, the number of cars increases dramatically, which makes it difficult tocontrol and manage the vehicular traffic. The emergence of the intelligent transortation systemrelieves the traffic pressure, and the number of the license plates is the "identity" sign of the vehicle,so the technology of the license plate recognition is an important component of the intelligenttransportation system.License plate recognition is based on the theory of image processing and pattern recognition.It locates the position of the license in the images by analyzing and processing the images of thevehicle license plates, and further more it extracts and identify s the characters. This paper studiesthe technology of the license plate recognition from three aspects, including the license platelocation, character segmentation and character recognition.Firstly, the paper studies the module of license plates.This module accomplishes the licenseplate location by adopting the license plate location algorithm based on color information andgeometric features. Secondly, the paper preprocesses the located license plate images. It adoptskinds of methods to process the license plate images, including color plate image grayscale, tiltcorrection, binarization, morphological processing, the removal of license plate frames and theprocess of non-character regions, and then obtains the binary license plate images only containingcharacter regions, which makes preparations for the characters segmentation. Once again, the paperstudies the character segmentation. It segments the seven characters of license plate by projectionalgorithm, and makes them have the same size with the characters in template library by normalize dprocessing, which lays a solid foundation for the character recognition. Finally, the segmentedlicense plate characters are recognized in the paper.It adopts the improved template matchingalgorithm to realize the characters identification, and introduces the implement of the algorithm indetail, and utilize s the MATLAB to simulate the experiments of the algorithm, the results show thatthe algorithm has a certain adaptability and robustness.In addition, this paper uses the MATLAB to simulate and implement the algorithm, and teststhe validity of the algorithm. The experiment shows that most of the vehicle images acheivesexpected results.