Research And Implemen3tation Of License Plate Recognition System Based On OpenCV

With the rapid development of Intelligent Transport System(ITS)based on computer vision,License Plate Recognition system becomes an important and hot topic of intelligent transpot system.The purpose of license plate recognition system is extracting useful text information of license plate taken from camera equipment.There are four main processes in the license plate recognition:license plate location method,license plate correction method,character segmentation method and character recognition method.License plate location method is the first process of license plate recognition system.The precision of location is very important to the performance of license plate recognition method.License plate location method is extensive attention of researcher.Although many relative algorithms were proposed,the precision of location needed to be further improved.For instance,license plate location method based on color is hard to process the vehicle image that the color of car is similar to that of license plate.The license plate correction method mainly overcomes the inclination of the license plate,so as to further improve the precision of the license plate recognition.The accuracy of character segmentation method is conducive to the improvement of character recognition.Character recognition method is as the last part of the license plate recognition system.Its accuracy directly affects the performance of the license plate recognition system.In this paper,the existing license plate location method,the license plate correction method,the character segmentation method and the character recognition method are analyzed in detail.The improvement of the license plate location method is mainly studied.For the rapidity of the license plate location method based on color and the validity of the license plate location method based on texture,a new approach of license plate location integrated texture and color is proposed.The main research results are as follows:(1)For the problems of the license plate location method,the license plate correction method,the character segmentation method and the character recognition method,the main researches in recent years are summarized and analyzed.Then,the problems and application area of these research are pointed out,which will further make clear the orientation of the license plate recognition system.(2)Because of the high computational complexity of the existing license plate location method combined color with texture,a novel method of license plate location based on color and texure is developed.First,the similarity between the pixel and the plate background is calculated.Then,the connected region of license plate is obtained by image morphology to complete the rough location of the license plate.Finally,the Sobel operator is employed to achieve the texture image of license plate,and the accurate location of the license plate is completed.(3)The development of license plate recognition system is completed.In other words,the license plate location method,the license plate correction,the character segmentation method and the character recognition method are realized.While the system is applied on the vehicle images collected from network,it can obtain a good precision and accuracy.In addition,compared with the license plate location method based on color or texture,the proposed method of the license plate location method could increase precision of 1%-12%and accuracy of 2%-12%.The proposed method obviously outperformed the license plate location method based on texture in terms of running time.