Research On The Algorithm Of License Plate Character Recognition

License plate number as the identity of vehicles, is the only evidence which atraffic management uses to do the vehicle violation penalties, charges, registration. Withthe development of society and improvement of living standards, the traffic more andmore developed, and the license plate recognition is the key to realize the intelligenttraffic management. Now, the license plate recognition technology has been widely usedin many fields of traffic management. As the increasingly high demand of the trafficmanagement, the further research on the license plate recognition technology isnecessary.This paper collects and analyzes related literature of license plate characterrecognition algorithm at home and abroad in recent years, and has an in-depth study andimprovement for good algorithms. According to the multiple characteristics of licenseplate and using image processing technology, there is a single character and then toidentify it. The main works of this paper are as follows:â‘ License plate characters segmentation: First of all, use image processingtechnology to do color vehicle license plate after location, such as image graying, imagebinarization, uniform background, median filter, tilt correction processing, removing thelicense plate frame. Then use a character segmentation algorithm which combined withthe connected domain and projection to do license plate image, extracting individualcharacter. Finally using normalization and thinning, the sizes of various characters isunified and the skeleton of each character is extracted, readying for extracting characterfeature of license plate character.â‘¡License plate characters recognition part: This part is the core part of the paper.On the basic analysis of the license plate recognition method commonly used, the paperproposes a kind of license plate character recognition algorithm based on SupportVector Machines. On the feature extraction, the algorithm combines the grid featurebased on number of pixels and peripheral feature; on classifier design, the algorithmadopts Support Vector Machines (SVM), and using SVM toolbox-LibSVM.This algorithm was achieved in Visual C++6.0programming environment. Theexperiments show that the algorithm is proposed for vehicle image positioning and character segmentation, has high accuracy, fast computation speed.