Research Of License Plate Number Recognition Method

With the continuous development of our country's economy,the number of cars increases sharply.Intelligent transportation system is becoming more and more important.So as one of the most important parts of intelligent transportation system,the study value and use value of the license plates recognition become much more vital than before.This thesis mainly conducts a study on the license plate number recognition algorithms.It mainly includes four parts: image preprocessing,the license plate positioning,the license plate segmentation and character recognition.Through the understanding of the characteristics of the license plates in China,this thesis has a profound research on the license plate number recognition algorithm.And it also analyses and compares the advantage and disadvantage of different algorithms.Moreover,some existing shortcomings of algorithm have been improved.So the algorithm has a better application in the process of license plates recognition.First of all,the acquisition of license plate image has been pretreated,including gray processing,edge detection,Binary.Then based on the mathematical morphology and projection algorithm,the license plate position has been completed.After the location of the mathematical morphology,the tilt of license plate is adjusted basing on the algorithm of Hough transformation.In the module of license plate character recognition,this thesis mainly does the researches of the advantage and disadvantage of character segmentation algorithm which is based on the study of template matching and makes improvements on its shortcomings.It also improves the establishment method of template base and the character matching method.Meanwhile,it also raises the accuracy of character recognition.The rate of identification of license characters is 92% and the time of identification of license characters is 500 ms in this thesis.In the process of the license plate segmenting,this thesis mainly analyses and compares the advantage and disadvantage of different algorithms including the license plate segmentation algorithm which is based on template and the combination algorithm which is based on the license plate projection and template.And it finally adopts the combination algorithm to segment the characters of license plate.Finally,adopting all the methods mentioned above in the system of vehicle license plate recognition,the operators establish a new operation interface.The result of every step in the process of vehicle license plate recognition can be showed and kept by it.