Design And Implementation Of Embedded Vehicle License Plate Recognition Terminal

With the rapid development of the society, the life quality of people is becoming more and more satisfying so that lots of families are rich enough to own one or more cars, which causes many issues in the field of communication and transportation. In order to solve these problems, intelligent transportation is proposed, and it is catching more and more attentions from countries and originations all over the world. While license plate recognition technology is an important branch of intelligent transportation, which can be deployed in many fields.The function of license plate recognition includes video collection, real-time license plate location, character segmentation and identification. Embedded license plate recognition terminal is convenient, flexible and cheap, it can adapt to different environment so that it is in utility in real life.The traditional license plate recognition terminal, which is a large computer, is rich, difficult to install and huge in size, all this features limit the development of itself. In this paper, an embedded license plate recognition terminal is designed to capture instant video and to display that video using Linux system on the base of study of V4L2 and FrameBuffer. Vehicle License Plate Location is realized using edge detection theory, then, character tilt correction and segmentation are done by combining principal component analysis with projection theory. On the base of character feature extraction and BP neural network, 37 different features are extracted in this paper, then, character classification and recognition are followed. What is more, simulation experiments are done to confirm that it is more perfect than traditional ones, and it’s possible to be used in reality.