Design And Implementation Of License Plate Recognition Based On Color Digital Image

Vehicle license plate recognition is a major component of modern intelligenttransportation systems, in traffic management, vehicle inspection, vehicle trafficmonitoring and other aspects it play a very important role. Generally, vehicle licenseplate recognition system consists of three parts, namely the subsystem of license platelocation, the subsystem of license plate character segmentation and the subsystem oflicense plate character recognition. In these three subsystems, subsystem of licenseplate location is most important, it directly related to the license plate charactersegmentation and character recognition success. The current vehicle license platelocation method is mainly divided into two categories, include the locate methodbased on gray-scale images and the locate method based on color images.Since the1990s, after years of research scholars at China and abroad, licenseplate recognition method based on gray-scale image is relatively mature. But in theapplication of this system, because of the camera image vulnerable to hardware,license plate using objective factors of time, light intensity, and the weather changes,the quality and characteristics of the vehicle license plate image area could producemany impossible to predict change, based on gray-scale images of vehicle licenseplate recognition system cannot be satisfactory results. In the above context, as wellas the processing speed of the computer is improve, the development of imagingtechniques, it is necessary to conduct further study on based on color image licenseplate recognition system.In this paper, the second chapter details the typical current license plate locationmethod, and then combined with the development of today’s cutting-edge technology,the proposed license plate location method based on color images, which mainly usethe license plate of color, texture and shape features, the experimental results show that this method has achieved good results, so this system achieved plate positioningsubsystem; then turn paper also describes the subsystem of license plate charactersegmentation and the subsystem of character recognition. Among the subsystem oflicense plate character segmentation, the first is use the method to achieve the Houghtilt correction on the license plate, and then further on the license plate of the binaryprocessing and removal of borders, the final is accurate segmentation of the charactersfrom the license plate; at the end of this topic also implements a subsystem of licenseplate character recognition, which uses a simple and relatively easy to implementtemplate matching method based on the experiment, and it has good results.Finally, statistics and summarizes the results from the use of matlab2010software on a highway toll station surveillance video to extract the low-quality imagesand high-quality images with the camera were collected for testing. Experimentalresults show that the realization of color images based license plate recognitionsystem, in a complex environment can achieve high positioning rate, and has strongstability and robustness.