Research And Realization Of Vehicle License Plate Location And Slant Correction

License plate recognition system is one of the most important components in the intelligent transportation system. It can be widely used in security and transportation field such as car parking charges, highway toll, traffic data collection, access control and prevention of crime. The system mainly divided into three subsystems, namely license plate location, character segmentation and character recognition. License plate location is the premise and foundation of character segmentation and character recognition, and its accuracy has a direct impact on that of the latter two aspects, thereby affecting that of the overall system. Therefore, researching the license plate location algorithm has a very big practicality.Firstly, this paper summaries the existing license plate location algorithms and analyzes their advantages and disadvantages, on this basis, taking it into account that the plate region has a rich edge information and color information, and the positioning method of only using edge detection is susceptibility to noise, the location method of simply using color feature is sensitive to body color and background color. This paper uses the method of integrating color feature and edge detection to locate the license plate.Secondly, this paper researches the slant correction algorithms. By now, the current two mainstream methods of detection angle are Hough transform and Radon transform. Comparative experiment is done, finding that the computational complexity of Radon transform is lower than that of Hough transform, the speed is faster than that of Hough transform, so this paper chooses the Radon transform to detect the tilt angle of the located license plate.Finally, this paper uses Microsoft Visual Studio 2008 programming environment to realize positioning and slant correction algorithms in C#. The actual 100 car images in the road and 500 vehicles images on the highway are choosed to test the system.The experimental results show that the method is better than single using one of the two methods and that its accuracy rate reaches 95%. Meanwhile, it has good real-time performance.