Video-based License Plate Recognition Technology

With the Supply and demand Contradictions ,which is Transport and transport facilities,is becoming more and more intensified, intelligent transportation system is considered to be a good way to resolve this contradiction. License plate recognition technology is the basis for intelligent transportation systems an important means of information acquisition, and more and more scientific research institutions make it to be their target of research. License plate license plate recognition system's difficulty is to. Locating part of the license plate accurately or not directly determine the license plate recognition's accuracy.In this paper, we studied a new type of license plate recognition system, this license plate recognition system's data source is the video data. In this system, with the different aspects of image acquisition, a dot matrix method is used for detecting vehicle license plate location. The vehicle license plate location of the system saves a lot of computing time and reduces the background interference for follow-up algorithm, and increases recognition accuracy of the system. This article also uses the system's using video image frame sequence and image source data of multiple frames to propose a quality feedback module, which can detect the result of a license plate recognition. We set up multiple evaluation parameters in the system during the middle of every espects and design the evaluation function.While the system recogned the license plate,the evaluation function uses the evaluation parameters to judge the accuracy of the result of recognition.If the accuracy below a set value,the system will go back to the link of selecting right image from the video image frame sequence. If not, the function believe the result is right. Since the vehicle license plate roughly location algorithm the article proposed before, the system's computation time is shorter and the background interference is reduced which can increase recognition accuracy to some extent.It can guareantee to add to system will increase little computation burden.In general, the complexer the algorithm's computation,the higher the algorithm's accuracy and the slower its computation speed. In this paper, we also put the proposed methods or algorithm compared with other traditional methods and algorithm.We found the way to achieve a reasonable balance of both of the system computation speed and the system's accuracy.