Research Of Digital Image Enhancement Prcessing And Data Minging Applications For QR Code

As a cheap and convenient mean of data transmission,QR code can be used by 020 E-commerce companies in storing and tracking order information and it can be easily decoded.Besides that the huge amount of orders' data that the enterprise has accumulated also need data mining technology to turn into a more intuitive operation index. But in the real world, the decoding of QR code can still easily failed because the poor image quality caused by scanning process and the environment. The shake of the camera during the scanning process will cause motion blur. The lens defocus will case the out-of-focus blur. The depressing light condition will lead to a low contrast image which fail the decoding processing. And the QR code too far away will be captured in low resolution which cannot provide enough information. This paper describes the basics of QR code encoding and it's error correction. The system implements adaptive histogram equalization to enhance the contrast of QR code image under weak light, and improve the performance of QR code recognition. The system combines Wiener deconvolution and Richardson-Lucy deconvolution to remove the out-of-focus blur, and the blind deconvolution method with LO norm based regularization to handle the motion blur, both methods reach a good peak signal-noise-ratio in restrictive runtime. The system use the gradient decent method to solve the robust formulation of super resolution problem which estimate the high resolution image from multi low resolution images and definitely improve the recognition rate. After solving that problem,this paper do the futher data mining and the data analysis on several indicators based on the mass and disordered data decoded from the QR codes.