The Research And Implementation Of The Character Recognition Of Vehicle Nameplate

With the economic development, purchases of vehicle increase sharply, resultingthat traffic problem is increasingly serious. Therefore, how regulate the trafficreasonably and effectively, has become the focus of today’s society. Intelligenttransportation systems can effectively solve the traffic problems and has now becomeone of the main methods of traffic management. This paper researched a corecomponent of the intelligent transportation systems, vehicle license characterrecognition technology(LPCR).LPCR technology, the efficiency and accuracy of which affect the performance ofthe entire recognition system, is the terminal part of the automatic license platerecognition system. At present, LPCR technology is widely used in various fields oftransportation and has extremely important practical value in terms of trafficmanagement, public safety and so on. In addition, LPCR is also one of the importantresearch projects in the transport sector.However, in different environmental conditions, taken license plate characters areeasily influenced by factors such as shadows, light intensity. License plate charactermay sometimes have the phenomenon of rupture, adhesions or fuzzy, that seriouslyinfluence the accuracy of the identification. In order to solve these problems and meetthe actual needs, the paper, combining computer technology and digital imageprocessing theory, based on template matching algorithm, put forward a mixedtemplate matching algorithm. Then, by use of the mixed programming technique ofVC++and Matlab, the paper designed a license plate character recognitionexperimental software.Specific research contents are as follows:First of all, the paper applied digital image processing theory to realize imagepreprocessing operation of license plate character images. The image preprocessingmainly included grayscale, image enhancement, binarization, image denoising andnormalized. Then, get a suitable normalized binary image for the license platecharacter recognition.Second, the paper researched the application of template matching arithmetic inLPCR. Summarized some common improved template matching algorithm in recentyears, and by the experimental verification, obtained the efficiency and accuracy ofidentification. And based on the improved algorithm, the paper also proposed animproved template matching algorithm, named as mixed template matchingalgorithm.Again, the paper applied the mixed programming technique of VC++and Matlabto design the license plate character recognition experimental software. In theprocess of software design, combining Matlab’s powerful tool for digital imageprocessing and VC++’s excellent interface design framework can improve the efficiency of program development.Finally, collect experimental samples of license plate character through the photofrom the practical scenario. And applicate license plate character recognitionexperimental software to recognize it. Experiments show that mixed templatematching algorithm proposed in this paper can effectively resist the environmentalinterference and improve recognition accuracy and efficiency.