The Research Of Recognition Algorithm Of QR Code And Its Application In New Access Control System

In recent years,with the popularity of smart phones and the expansion of electronic payment,two-dimensional code appears everywhere in life.As an important member of the two-dimensional code,Quick Response(QR)code has been applied in many fields,such as electronic commerce,warehouse management,item tracing,hyperlink of website and so on.In this paper,the recognition algorithm of QR code and its application in a new type of access control system are studied.The QR code recognition algorithm mainly consists of four steps:image preprocessing,location of QR code,distortion correction and data decoding.In this paper,we research the preprocessing algorithms of QR code,such as extract the gray image,image filtering and binaryzation.For the Otsu algorithm is not suitable for the condition that the light is not balanced,this paper studies the binaryzation algorithms that get the threshold in local image.This paper also studies the location algorithm of QR code.The algorithms based on Hough transformation and HOG feature of the image edge are studied.But the two algorithms are easily affected by the background.Therefore,this paper studies the traditional algorithm that based on Pattern Finder,and optimizes this algorithm,which reduces the amount of computation and improves the speed of algorithm.This paper focuses on the correction algorithm of QR code.A method for the distortion classification of two dimensional codes based on the linear characteristics of the four edge points of a QR code is proposed.For the QR codes with squint distortion,this paper detects three comers of QR code through Pattern Finders,and then uses the edge lines to get the fourth corner points,and finally corrects it by perspective transformation.Because the algorithm of relational mapping can not ideally correct those complex distorted QR codes,a method based on dynamic curve fitting is proposed to mesh the complex twisted QR codes.The algorithm proposed in this paper finds key points in the edge of module based on detection lines,and then to get the parameters of edge curve by curve fitting.It is easy to extract original data from twisted QR code with the help of sampling curves.The experimental results show that the method proposed in this paper can effectively read the data in the distorted QR code.Finally,this paper establishes the device of access control system based on Hi3518,and applies the recognition algorithm of QR code proposed by this paper to the access control system,and tests the QR code with different material and distortion.The experimental results show that the access control system can effectively recognize the two-dimensional code and realize the other functions.