Research Of The Vehicle Licence Plate Recognition System

In recent years, with the rapid development of economic, a substantial increase in thenumber of cars, the traditional manual management methods obviously can not keep therhythm. For this purpose, Intelligent Transportation Systems (ITS) is currently the maindirection of development of traffic management, and is also the cutting-edge topics andresearch focus of competing research and development in the transport sector of countriesaround the world. However, License Plate Recognition (LPR) System is one of importantpart and key technology of Intelligent Transportation Systems. LPR is a computer visionsystem, which can recognize the numbers of license plate automatically, and can be used inautomated highway tolling system, parking management system, stolen vehicleidentification system, guard system, intelligent transportation system and so on. The widerange of applications in LPR will help to speed up the process of traffic managementautomation in our country.This paper made a comprehensive discussion of images preprocessing, license platelocalization, character segmentation and character recognition and other aspects related totechnology, algorithm and system overall design in the License Plate Recognition Systems.Image preprocessing: It uses the methods of gray level stretching, contrast stretching,image smoothing, edge detection, image binarization and others to do image preprocessing.And then the quality of the image is improved, the image area is strengthened.License plate location: This section is a crucial step in the License Plate RecognitionSystem. This paper takes the image based on the edge point, and it locates the license plateimage according to the vertical texture feature of license plate region. This method can givea satisfactory recognition effect.Character segmentation and recognition: First, the method of inclination rectification isapplied to license plate image. Second, the corresponding connection region is used todivide up and get the isolate character. Third, the characters are normalized, preparing thecharacters of the same size for the next character recognition. And then in characterrecognition section, we choose feature extraction method with strong adaptability, selectmany feature points, and combine with the template matching method, to identify the character of Chinese characters, letters and numbers quickly, a high recognition rate andreliability are obtained.On the basis of the above study, we make many pieces of car images collected as theexperimental sample data, use VC++6.0to develop the car license plate location and therecognition system software. Experimental data shows, this system can locate and divide upthe car license plate, divide up and identify the character. The system developed is suitablefor China’s current development status of the transport vehicle industry, has good marketvalue for development. At the same time, the system involves the field of digital imageprocessing, pattern recognition and other technologies.