Research On Denoise And Edge Detection Algorithm In License Plate Recognition System

With the development of economy,the automobile has become the common means of transport commonly used household,how to effectively solve the enormous pressure of transportation management become the most critical issue.With the increase of informatization degree,image information has become an important source of information,license plate recognition(LPR)technology becomes one of the core technology in intelligent traffic management.License plate recognition technology is based on the digital image processing technology and the recognition technology to identify the license plate number and then realize intelligent traffic management.In traffic monitoring,highway management,violate compasses record,the park management and other fields,LPR is widely used for traffic management.Influenced by viewing angle,scale,brightness,stains,vague uncertain factors,Vehicle images in the acquisition process is disturbed by the noise,subsequently,the images which the quality has been lower further affect the license plate edge detection,segmentation,recognition process.How to remove the interference to improve the vehicle image quality and to ensure the precision of edge detection and the subsequent process effective implementation becomes urgent problems to be solved in the license plate recognition technology.In this paper,we have emphasis the research on multi-noisen vehicle image de-noising and edge detection method.Firstly,we presented the theory of image noise,then mainly analyze the characteristics of the vehicle image noise,for single noise vehicle image and multi-noise vehicle image which be in accord with the characteristics of the vehicle image noise,analogy analyzing kinds of commonly used classical denoising methods by evaluation function,such as median filter.Secondly,for multi-noise vehicle image,we systematically presented and analyzed the theory of edge such as Canny.Finally,we have designed a kind of improved edge extraction algorithm based on median and mean filter to promote the quality and the accuracy of the image for the multi-noise vehicle image.Through the Matalab simulation,we have observed the PSNR(peak signal to noise ratio)and NMSE(Canonical mean-square error value),by analyzing the functionality of the proposed algorithms,we draw the conclusion:the method in this paper has been proved to be better performances than other methods in denoising and edge detection for multi-noise vehicle image.After verification,the algorithm proposed in this paper is proved to be certain practical value and theoretical significance.