Research And Accomplish On Technique Of Recognition License Plate

License Plate Recognition is the core techniques for the Intelligent Transportation Systems. It takes an important role in information collection, Vehicle Monitor System, Automatic Fare Collection System and other fields. By virtue of its simple equipments and economical equipment maintenance, The Vehicle License Recognition based on image processing is the present research focus on the Intelligent Transportation Systems.All color images of car licenses in the paper were acquired from web scanning and CCD cameras in a highway station at different time, different points of view, and different places. The difficulties incurring in the License Plate Recognition, such as uniform-illuminated, distorted Vehicle License Image, character-touching, character-broken number plates and so on, make our research significance and practical applications.The whole process of the system can be divided into three stages: license plate location, character segmentation and character recognition. In order to eliminate the noise and improve the quality of image, pre-processing of original images such as Image Enhancement, Filter Preprocessor, edge detection is needed before license plate location. Aimed at the difficulty in locating vehicle plate, an algorithm of vehicle plate location based on gradient projection together with color area segmentation is put forward in the paper. This algorithm is fast and its accuracy is high. In order to obtain gradient angle for emending the image, we present a method of random lines to detect edges of the vehicle plates. In comparison with the method of Hough transform, our method cause few storage and calculation amount, short calculation time. An algorithm, derived from the combination of the global and local image binarization methods, is put forward after analyzing the advantages and disadvantages of them in this paper. Our algorithm can not only retain image's details, but also eliminate disturbances of artifact. During the course of character segmentation, we make the best of the apriority knowledge of the normal license plate and use the vertical projection to segment it into single characters. Finally we use template matching to recognize the characters.We have realized the three stages of license plate recognition by using computer. The experimental result that this system can accurately achieve the three stage operation, especially for the alphabetic and figure character recognition.