Research On Automatic Recognition Technique Of License Plate Based On K-L Transform

The Vehicle License Plate Recognition (LPR) technique is one of the key part of modern Intelligent Transportation System(ITS). It can segment the license plate region and character automatically, and then recognizes the character. This thesis mainly research into three key portions of LPR.The first portion is location and segmentation of the license plate region. The location of license plate region is one of the important factors which exert an influence on LPR system performance. It affects the accuracy of character recognition directly. According to license plate own feature,the author locate and segment license plate region based on colour two-dimension histogram.The second portion is the preprocessing of license plate region which has been segmented. The purpose of license plate image preprocessing is to raise the speed of character recognition. This part includes binary processing, character segmentation, character normalization, opening filter based on mathematical morphology, etc.The third portion is the character feature extraction, and recognise the character. This part is the most important part of LPR. The author extract the feature of character based on K-L transform.