The Research On Automatic Vehicle-License-Plate Recognition System Based On Morphology

Automatic license plate identification plays an important role in Intelligent traffic control system, parking lot monitor system and automatic charging system .Vision-based approach is promising since it requires no pavement adjustments and is more flexible and economical.A Vision-based license plate identification system has to be robust, since the plate may be inclined or distorted, the characters may be blurred and the strokes become disconnected. The algorithms also have to be adaptable since the light condition and image capture position could vary.The license plate identification system has to implement license plate region identification, character extraction and character recognition .We use edge features and shape features to locate the plate region, in this process image enhance and mathematical morphology are used. In character segmentation the image is binarized and the style of license plate is confirmed, and the frame of license plate is cut off. Then, the image of a whole license plate is segmented based on pixel projection into several images, one of which contains only one character.These features are computed by analysis of concurrence matrix and project histogram. A modified Hough transform is also used in plate inclination detection. In character recognition process, Low dimension template matching is a robust algorithm since it works on highly blurred Images. Structure information is used to distinguish some alike characters.