License Plate Location Based On The Top-Hat Transform

With the high-speed development of our economy, our automobile industry has increased sharp. To ensure the traffic safety and improve economic efficiency, the Intelligent Transportation Systems has already become the main direction of present traffic administration development. The License Plate Recognition is the important component of the Intelligent Transportation Systems, and is useful in many fields. The system can automatically register, verify, monitor vehicle with the automatic recognition for vehicle license plates. So it can be used in many kinds of occasions, such as the charges system of expressway, monitoring system at road and roll-gate, charge and monitoring system at the district, parking area, guide system, the system counting quantity of vehicle passing in a certain period time, etc.Generally, LPR system mainly comprises modules of image acquisition and preprocessing,plate location,character isolation and character recognition. With the development of the optical technology, Currently, it is easy to get clear and vivid vehicle image in all kinds of application conditions. So the plate location technology is the important foundation of the follow-up work. Research about the plate location are the focus of the license plate recognition technology ,and many algorithm has been proposed.Two key indicators of license plate location are accuracy and speed of the algorithm implementation. As development of computer hardware technology, microelectronics, and embedded technology, Recognition speed has greatly improved. The main problems of License plate location is how to improve the recognition accuracy in different environments such as weather conditions, changes in light intensity and background noise interference. More that,license Plate itself can be tilted, soiled, and other exceptions for long time.The kernel of our paper is how to eliminate the interference and determine the location of the license plate quickly, and ensures the accuracy. Learning much of algorithm on LPR, we propose a new plate location algorithm based on Top-Hat to find the area containing the plate, and then relocate by the projection.In the current field of computer vision, mathematical morphology receives more and more widespread attention and vigorous research by it unique ideas and effective practice. Mathematical morphology is an effective non-linear theory for image processing, With its complete concept, method and algorithm system. The basic idea is measuring image and extracting target by using certain forms of structural elements. The main content is to design a set of concepts and operational methods of transformation, to describe the image of the basic characteristics and the basic structure, which is the image pixel set, or part of the mandated Sets Relations. The fact shows that mathematicalMorphology is essentially a nonlinear filter, and widely used in image filtering. Many scholars have proposed a series of morphological filtering algrithments for practical application. Edge information can show characteristics of the objective, and is important to descript image, being used to image classification, recognition and retrieval. So Mathematical Morphology has a wide application in Image feature extraction, object recognition. Usually we construct texture feature in vector by mathematical morphology operations, and then classify texture in order to achieve the purpose of texture recognition.Compared with traditional mathematical morphology application in grayscale images focusing on expansion, and corrosion, top-hat transformation can eliminate the background noise effectively. Based on the experimental image features in our experiment, repeat experiments to construct the most appropriate structure element. In addition, for the complex environment that often appear in pseudo-plate region, we propose character-based projection for the second positioning location, can obtain license plates very accurately.