License Plate Location And Character Segmentation Of LPR System

LPR(License Plate Recognition) System is a practical image processing and recognition system which is of very ample application prospect, important realistic significance and highest economic value.At first the formation, principle and research survey of LPR system is introduced, while the purpose and significance is pointed out; and then all sorts of characters of license plates are discussed in details, and all kinds of existing LPR methods are deeply analyzed. On the basis of the preceding work, a sequence of image efficient algorithms for image segmentation process are put forward referring to license plate location and characters segmentation, which are the two keyprocedures in LPR System.The main content of this thesis include:1 .License plate location: On the grounds of the feature that license plate area contains edge information of high intensity and high frequency, a locating algorithm based on edge analysis and texture information statistics. Finally, the algorithm is improved according to the problem arising in the test.2.Preliminary process of license plate image and ideal characters segmentation: At first, image preliminary processing is discussed among which threshold selection is highlighted and a image thresholding algorithm based on improved Bernsen Method; then the concept of ideal characters is introduced, and a algorithm for ideal characters segmentation based on vertical projection is presented.3.Ordinary characters segmentation: Ordinary characters is discussed in details and classified reasonably. Four algorithms is put forward referring to the features of all kinds of ordinary characters.4.Programming and analysis of the experimental results: The algorithms which are presented in the thesis are programmed, and in the mean time, the experimental results is analyzed concisely.