Application Of QR Code Recognition Algorithms On Smartphone

This article mainly describes the process of recognising QRCode image and making image gray scale, median filter, binary image. Then barcode recognition algorithm process the image got from smartphone and restore the data from image matrix according to the specification of QR Code. And this article also describes the application of recognizing QR Code image on smartphone, such as message, wap and vcard recognition.The images which captured from smartphone are influenced by blamed shooting environment, shooting angle, curl image and noise from camera. Some of them are a bit tilted or blurred. Another restriction is set by smartphones themselves, for those handset devices have very low computing ability and have strict demand on memory usage. So they require high quality preprocess, our barcode recognition algorithm requires small time complexity and small storage complexity.Based on above problems,we adopt gray scale technology after we get capture the image and this technology not only reduces storage space but also is convenient for histogram technology. And then we adopt median filter technology, it is easy to wipe off isolated point, noise of line and can keep the edge of the image. At last we adop local binary image algorithm and these preprocess solve the problem of blurred image. Based on above problems, the implementation method described in this article utilizes methods for noise cleaning and image correcting. Because the bar code symbols are printed as two color images and the main noise influence quality of image is called Salt-Pepper noise, applying median filter gives satisfying result. For image threshold, because images got from camera is very blurry; threshold method based on local threshold value satisfied the demand of next image processing request.The implementation method described in this article utilizes a grid sampling method for QRCode image data sampling, this method makes a sampling grid according to the configuration of image, build a grid matrix independent of image itself, the test result shows it is a proper way to sampling data from QRCode image.