Research And Implementation Of License Plate Recognition Method Based On Image Processing Technology

With the development of society, Our country’s auto industry gradually moves toward prosperity from weak to strong. Meanwhile, it also brings enormous challenges: a sharp increase in road traffic, urban congestion, traffic safety, and many other problems. To make our city traffic more smoothly, travel more convenient, intelligent transportation system(Intelligence Transport System, ITS) get more and more people’s attention. License Plate Recognition System(License Plate Recognition System, LPRS) is one of the core technology of intelligent transportation System, and is also important in pattern Recognition and intelligent control technology, which relates to digital optics, digital image processing, computer vision, and other fields, and which is widely used in highway automatic charge(Electronic Toll Collection, ETC.), intelligent parking lot management, road traffic statistics, as well as Electronic police, and so on. Therefore, further study of license plate recognition system has an important practical significance.In this paper, license plate recognition system is made up of four parts, namely, license plate extraction, license plate skew detection and correction, character segmentation and character recognition algorithms. Not only learn the commonly used image processing methods and further study of the license plate recognition algorithms, but also put forward some effective improvement measures. On this basis, license plate recognition system was developed. The main content of this paper consists of the following components:(1) This paper introduces the research status of license plate recognition system at home and abroad and the significance of research, and specifically points out that our country faces the technical difficulties in the license plate recognition technology.(2) In the license plate localization algorithm introduces the commonly used image processing methods, image processing for further, to covert color image to a gray image first, which can improve the system operation speed, shorten the operation time. For locating license plate, first edge extraction using edge detection operator, obtained through the comparative analysis of Robert and Sobel operators, improved after Robert edge detection operator has more superior performance. Then, image morphological opening and closing operation can be used to locate the license plate location.(3) The license plate image captured may not just level, we need to test at this time. Hough transform and Radon transform is introduced to measure the tilt, and algorithm implementation Hough transform with better effect. Subsequently, to correct the tilt plate operations, obtain a horizontal image.(4) For complete license plate location and license plate images of the skew detection and correction is executed, base on prior knowledge and distribution of the license, we’ll have a segmentation of the characters. To make it the same as the characters in a template library size and convenient for character recognition, then character normalization may be done split operation. Character recognition is the normalized characters segmentation of character in contrast to match the characters in the template library circulation and the output is similar to most characters..(5) In this paper we study the relevant license plate recognition algorithm and explain the development of related processes and procedures, using the MATLAB software developed a suite of license plate recognition system.