基于OPENCV的车牌识别设计

AbstractWith the increasing popularity of cars,traffic congestion andaccidents may often occur.The signification of car plate like residentidentification card,that was the unique identification,so the meaning ofthe ability of vehicle management was important in solving the above-mentioned issues.It is obviously difficult to recognize the license plateonly by manual means.In this regard,the technology about automaticclassification of license tag number is of major value to solving theproblem.The thesis proposes a license plate recognition method and itbased on OPENCV,then I will concretely analyze and study the existingrecognition method.The specific research contents include the following:(1)License plate location:this paper adopts grayscale andbinarization.In position part,HSV color space is used to realizing thecoarse positioning of license plate via the color features of license tagimage.(2)Character segmentation:binarized the license plate picture,scanning each column of pixels,the left and the right boundary of thecharacters can be determined by the discrepancy of pixels with the colorof black and white,and divided the characters after normalization,to fullyprepared character recognition.(3)Character recognition:Support vector machine method was founda hyperplane in the trait space of the image.This hyperplane can be dividedinto many different categories.The most important step is to process thetest sample to obtain the characteristics,and found the characteristics'distinction in the test image,finally using the SVM classification todistinguishing it.Key words:OpenCvHSVImage RecognitionCharacterSegmentationLicense Plate Location