The Research On The Quick Response Code Image Preprocessing

QR code has been applied to the train service management during the 2010 Spring Festival,which indicates that the domestic application of QR code has begun to be widespread. QR code recognition was placed in a high position。Now, through the QR code's acquisition and recognition often require highly specialized equipment. So the convenient, efficient devices,such as the method using usual camera and image processing, come into being,and will develop into the mainstream of the future.The process which used image processing methods to identify QR Code symbol can be divided into:image acquisition,image preprocessing,image orientation,image rotation,sample image,image decoding.QR code symbol acquisition with the camera will be generated noise such as hands shake, brightness of light, optical collection system performance, the scratch on the surface of bar code image, and other reasons.In order to ensure a better recognition rate,the QR code image must be collected by a certain degree of pretreatment.The usual pretreatment steps are gray level transformation, filtering, binarization. Through detailed analysis on the process of filtering and binarization,pretreatment steps is proposed: gray level transformation, noise removal, binarization, median filtering. The main difference between these two steps are as follows:(1)The sequence of pretreatment steps is different.This paper proposes different method from the conventional method,stepped binary on the front step of median filter, and a suitable binary image median filtering algorithm is proposed. the effectiveness of this approach is proved.(2)To decrease the salt and pepper noise in the QR code images'collected,median filter coefficients is proposed . And a set of coefficients are found,which is used in the experiment to increase the QR code image recognition rate. (3)This paper considers the Gaussian noise in the QR code image denoising case, the wavelet denoising applied to image denoising QR codes, the analysis of various denoising effect, and propose a new wavelet thresholding algorithm, simulation results proved that the algorithm used in the QR code image denoising can improve the efficiency and accuracy of decoding.(4)On the QR code image captured by camera often has uneven and reflective light phenomena, a dedicated QR code image binarization method is proposed, results show that,the recognition rate using this method than the commonly used global threshold 2 value method to high recognition rate than the local binarization method fast.This thesis constitutes the above findings into a complete set of preprocessing steps , firstly,gray level transformation , secondly,noise removal , once again ,binarization;and finally median filtering.Last,this article will take one patent of my invention as a case to attest——Use two-dimensional code of mobile to replace the method and system of bank card payment function.