License Plate Intelligent Recognition System For The System Of Expressway Toll Collection

The License Plate Intelligent Recognition Sytem (LPIRS) is the artificial intelligent system which processes the collected digital images containing license plate, automatically locates the position of license plate, and realizes the characters seperation on the plate. Therefore, the LPIRS consists of digital image collection and pre-procession, characters seperation, extracting characteristic and recognizing characters.Based on the binary image, the thesis proceeds a suit of algorithm for the LPIRS serving the the Expressway Toll Collection System (ETCS), drawing on the experience of the thought of seeking the area of plate based on the grey image.The method of setting the location of the device for image collection could remove the influence of the device's location. Hence, the algorithm attains the objective of rectifying the angle of inclination firstly.Directed against the scene in special where the image is gathered, during the procedure of pre-processing, the algorithm of first difference combining the method of fixing the threshould of binarization via accumulating the points is used to complete the job of extract the outline of the image.Taking advantage of the horizontal texture characteristic and the physical trait in the plate area, the algorithm locates the position of the plate roughly, then improves the pricision of location in the area, and ascertains the area finally during the plate area location.The algorithm designes different method of characteristic extration and recognition for the two. For the Chinese characters, the method of nonuniformly rough mesh and the second-order moment is used for characteristic extraction, and the recognition method of BP neural network enhanced is also used. The density per unit combining the origin moment is used for extracting the characteristic of numbers and letters, avail of the method of dynamic clustering.