Vehicle License Plate Recognition Technology To Achieve

Vehicle plate recognition system is a research project related to multiple research fields including image processing, pattern recognition, computer vision, neural networks and software engineering and so on and has large market and promising prospect of commercial use.In this article, we use mathematical morphology methods to locate the vehicle plates and template matching to recognize the characters on the plates.To locate the vehicle plates, we first commit preprocessing to the images as to focus on the information of the region in which the plate locates while weakening other regions. The preprocessed image will be easier for recognition process. The jumping scanning is used to define the upper and lower boundaries of the processed image and the vertical projection is used to define the left and right boundaries. After the process, the plate region is coarse positioned and secondary locate process is called for. The obtained region is small enough to get precise threshold value to separate the background and characters the thus to attain the accurate plate region.The second step is to divide the characters. First, we commit Ostu binaryzation method to the located vehicle plate region and use vertical projection to divide the characters. The region with the largest space between the characters can be determined to be the region in which the second and the third characters locate and then divide each character based on the internal characteristics of the vehicle plate.The template matching method is used in the recognition process. The template is constructed as follow steps: (1) Obtain eigenvector based on the specific matrix of the character division; (2) Find out the distance between the eigenvector of the object character and the eigenvector of the template and determine the character through the distance. The determine process can be minimum distance method and k-neighbor method. The system introduced in the article is used to recognize more than 1000 images contained vehicle templates, the correct rate of location and recognition are both more than 86% which means that the result is satisfied.