| As the number of private vehicles increases rapidly,speeding vehicles, vehicles theft,overloading and other kinds of illegal behaviors have become more and more serious. The vehicle license is the only vehicle identity. So the scope of the application of the vehicle license plate recognition system has been greatly expanded.The paper incldes: the red light violation detection and the recored of all passing vehicles and real-time blacklist access and alarm function and vehicle license plate automatic recognition function; illegal identification and judgment and punishment function, which can automatically identify whether the vehicle travels inaccordance with the rules of the road driving, and it records the traffic rule offense evidences. The system has realized the information query function, which can store the detected information in the database, then provide query function,to help the users to input a variety of query terms for information and analysis. After the system requirement analysis and design, the system has been deployed to operate, which,judging from the current running state,is good. After tests and experiments, the system can detect the vehicle information and carry out the information transmission and maintenance.The innovation of this paper is that: the clustering algorithm of the existing license plate recognition, K means, C means clustering algorithm of the license plate image recognition in the field of software engineering has great advantages in application. However, the number of the clustering center generated by the algorithm is too large, and the selection of the center point will lead to the problem of local optimal solution. To solve the two problems, this paper,based on the relevant theory of genetic algorithm uses a dual genetic algorithm to optimize the selection process of the center point, and proposes a DGC means fuzzy clustering algorithm. In this algorithm, the first genetic algorithm is used to optimize the clustering center. The genetic algorithm is used to determine the number of centers. Secondly, the algorithm takes the individual as the center of the cluster. |
