Design And Implementation Of Laser Reproduced QR Code Recognition System

The traditional anti-counterfeiting technology increasingly face a challenge, so it isparticularly important to study a more effective means. Hiding the QR code symbol on theanti-fake marking then using the laser holographic technology to identify it is a powerfulmeans of security. The hidden symbol can contain more information and can be quicklyrecognized. It can greatly improve the anti-fake performance and increase practicability.So it is important to research the technology of recognizing laser reproduced QR code.However, compared with the standard of QR code image, the laser reproduced QRcode image is blurred. It is not easy to reproduce a stable image and recognize it. In orderto solve these problems, in this paper, an embedded recognition system based on theAndroid platform is proposed. Research is mainly focused on the recognition method offuzzy QR code under the complex background, then an effective Laser Reproduced QRCode Recognition algorithm(LRQRCR) is proposed, and the whole recognition system isrealized on the Android platform. In the identification algorithm (LRQRCR), firstly, ruleof maximum is used to gray the original image and then median filtering algorithm is usedto filter the noise, so that more details is retained. The image is divided into pieces andthen Otsu’s method is used to operate each piece to create a binary image, so the problemof uneven illumination is solved. Secondly, effective area of QR code is separated frombackground by adopting region growing, then Roberts operator is used to extract the edgeof effective area. After that Hough transform is used to get angular points. Then a standardsquare image can be gained after the binary image is rectified by using perspectivetransformation and bilinear interpolation. Finally, a method of calibrating fuzzy QR codeimage is proposed to make the image clear, then it can be recognized.In this paper, a large number of related experiments have be done to validate thesystem recognition rate. Experiments demonstrate that majority of the original images canbe made into clear images and be recognized by the algorithm proposed. Due to the faulttolerant rate of QR code itself, it can be correctly recognized even if there is a smallamount of error. If the error is large, it can be recognized via similarity degree. Theexperiments have verified the feasibility of the recognition system.