Research On Plate License Recognition System Based On MATLAB

In recent years, the development of intelligent transportation has become more and more important. As an important aspect in intelligent transportation, plate license recognition system has taken more and more attention. The plate license recognition system can be applied to public parking, highway speeding management system, crossing road, district vehicle management system, and so on. Although now there are already some exsiting plate lecense recognition systems, the research and development of arithmetic have never stopped, and there are still many scholars who are doing further research and improvement.Firstly, the paper gives a deep research on the status and technique of the plate license recognition system. On the basis of research, a solution of plate license recognition system is proposed, and the paper focused on the software part. The whole system concludes three modules. They are plate location, plate character segmentation, and plate character recognition. In the plate location module, the paper puts forward an arithmetic of plate edge recognition by wavelet decomposition, and an arithmetic of locating twice, which improve the accuracy in bad light condition, and are fit for plates with different grounding. An improved Otsu arithmetic is used in the process of binaryzating, which reduces the running time, and can achieve good effect for different kinds of plate. In character recognition part, with the momentum of the gradient descent method, the BP neural network can fast convergence.Compared the BP neural network with template matching arithmetic, which improves that the BP neural network are better than the template matching arithmetic. Then, a test platform has been built with MATLAB, for the test of the system. Through the test of 353 monitoring car photographs, the results shows that the system can effectively meets the requirement, and lay a good foundation of technology for productization.