Research On License Plate Recognition Based On Binarization Algorithm

License plate recognition technology is the core of intelligent transportation system, intelligent transportation has become the main direction of the development for modern traffic management. In reality, because the environmental constraints, such as light irradiation uneven, license plate area is polluted or license plate character part such as incomplete, there are a large number of noise and texture background in license plate images. Through the computer digital processing of license plate image, and the accurate recognition of license plate character, we can not only obtain the relevant information of the car and the road, but also largely solve the problems of work intensity and statistical complexity and low efficiency by artificial recognition. Therefore, it is important to research and develop the automatic technology of license plat recognition.License plate recognition uses the automatic identification technology based on card image. License plate localization and binarization is an essential step in license plate recognition, because it influences directly the accuracy of character’s segmentation and recognition, plays an important role in the recognition system.A local and dynamic threshold algorithm based on LOG operator and a global value based on OSTU is proposed in this paper,which bases on the effective integration and discards the disadvantages. According to the noise character of license plate image, eight neighborhood and connected region detection method based on structural analysis is used to remove the noise in the image of license plate binarization. During the task of locating license plate,a method based on multi characteristics and multi methods is proposed, and achieves accurate license plate location; Experimental results of different occasions demonstrate that proposed method is able to achieve high recognition rate.