Car License Plate Localization Using Hole Filling and Support Vector Machine Approach

In this paper a novel technique is proposed for detecting the location of The License Plate (LP), LP position is detected by examining an image of the car or a frame of a video sequence. The image is segmented by using frequency filters. Wavelet analysis will be used to extract the image frequencies; the LP contains areas with high frequencies. Basically, it is a bright plate put on a darker color which is the car body color. The plate itself contains abrupt changes in color due to the existence of letters and numbers. Our idea is to find the regions with holes – which mean a border surrounding bright color, then put all regions with hole into LP pool, the exact LP will be detected according to its texture properties. The proposed algorithm showed high accuracy (97%) in detecting the position of LP.