Research And Implementation Of QR Code Recognition Technology Based On Android

Two dimensional barcode contribute to an evolution of automatic identification technology. QR code, which is wildly used in China, has many advantages such as efficient coding ability and quick identifying characteristics. Smartphone industry has developed rapidly in recent years. Android grabbed most of the smartphone market, as it’s open, extensible, user-friendly and easy to develop.The paper conducted an intensive study on QR code identification technology based on Android, includes following aspects.1、The images collected usually influenced by unbalanced illumination or defocusing blurring, which increased the difficulty of identification. The paper analyzed the images under abnormal illumination and adopted a light-balancing algorithm to adjust the gray value of each pixel. Experiments show that this algorithm can improve the abnormal illumination. In order to reduce the defocusing blurring, we set the disk function as point-spread function (PSF), using Wiener filter to restore image. Experiments on real data demonstrate that the method is effective.2、Exact localization is the premise and crux of QR code recognition, the paper put forward a locating method based on contour tracking and contour sifting, which can get three vertexes’positions of the QR code rapidly and accurately, then get the fourth vertex’s positions by approximating the edges. We estimate the version and rotation of QR codes using four vertexes’positions and the contours of the Position Detection Pattern. Experiments show that the algorithm can get the accurate location, affected little by the background of QR codes, rotation, distortion, etc.3、When the QR codes are perspective distorted or nonlinear distorted, they can’t be recognized directly. Perspective transformation and bilinear interpolation algorithm are taken for correction. We use block correction for the QR code with high version, which improves the calibration effects. Nonlinear distortion is complex, the paper proposed a method which can get the edges of QR code by morphological close operation and least square method (LSM), then correct the QR code image line-by-line. Experiment results show that the effect is obvious.4、On the basis of the image preprocessing algorithm, correction algorithm and other method proposed in the paper, we designed and implemented a QR code recognition system based on Android, which includes the UI display, image acquisition and QR code identification. The performance of the system is tested, optimization is proposed and the speed is improved as much as possible. Experiments show that the system can recognize QR code in a complex environment.