Research On Correction Algorithms For QR Code Image In The Complex Environment

As a new electronic payment media in various fields,the recognition technologies of two dimensional code has become the focus in automatic identification field,some domestic and foreign cities are beginning to apply two dimensional code to the subway ticket selling system.Nowadays,the two-dimensional code has many types,quick response(QR)code with high speed of reading,large storage and efficient error correcting has caused wide public concern.However,in the process of QR image acquisition,it is difficult to avoid some noises such as uneven illumination,tilt,or geometric distortion,etc.which may bring some errors to QR decoding.Therefore,some appropriate image restorations and corrections must be done.In this paper,coding principle and basic recognition flow of QR code are briefly explained.Moreover,the involved image processing algorithms are analyzed and compared.The main contents of this paper are as follows:Firstly,the structure feature,coding principle and basic flow of the QR code are discussed in detail,and its recognition process is divided into two parts: image processing and decoding.Then,the QR code image processing of each phase are described and simulated,and the results are analyzed and compared.Meanwhile,the paper analyzes some damaged QR codes which are collected in complex environment,and proposes an improved algorithm of adaptive brightness binarization based on image processing.The algorithm solves the correcting and recovering problem of QR code image under illumination unevenness.An adaptive median filter is proposed to improve the traditional Canny operator,furthermore,the straight line is identified by combining with Hough transform,and the tilt angle of the image is obtained to realize the tilt correction.Moreover,in order to achieve fast positioning and correction of the image,it is proposed to locate the three corner coordinates of the QR code by finding the position detection pattern.It overcomes the limitations such as burdensome and time-consuming of Hough transform method.Ensuing,the distortion correction control point is selected by using the projection transformation.And then the entire distortion correction process is completed through applying the bilinear interpolation algorithm to gray-level interpolation.Finally,the processing and correction algorithms,coding of QR code image are compiled to a system based on MATLAB software.Decoding and other functions are also integrated into the system to complete the GUI interface design.It makes the entire process visualization and displays the intuitive effects of various algorithms.