复杂环境下QR码图像的校正算法研究

发布时间：2019-05-22 16:29

【摘要】：二维码作为新兴的电子支付媒介在各个行业领域受到普遍关注,国内外一些发达城市开始尝试将二维码支付技术引入到地铁售票系统中,二维码识别技术也随之成为自动识别研究领域的焦点问题。发展至今,二维码的种类层出不穷,其中QR码凭借它在识读速度、存储容量、纠错能力等方面的诸多优点得到了人们更多认可,但是在QR码图像采集过程中难免会引入一些噪声如图像光照不均、倾斜或几何形变等问题可能会给QR码的解码带来一定程度的误差。因此,我们要对此类问题进行适当的图像恢复校正。本文对QR码识别的算法原理、基本流程做了简要说明,并对涉及到的图像处理算法进行了仿真和结果的对比分析。本文的主要工作内容如下:首先对QR码的结构特点、编码原理及识别算法的基本流程进行了详细论述,并将QR码的识别过程分为图像处理和译码两大部分。然后对QR码图像处理过程中的各个环节进行了阐述并仿真,后将实验结果进行分析与比较。同时,针对复杂环境下QR码受损类型进行分析,在图像处理的基础上提出了一种自适应亮度的二值化改进算法,解决了光照不均下QR码图像的校正恢复问题;接着提出了一种采用自适应中值滤波对传统的Canny算子边缘检测法进行改进,再将其与Hough变换相结合进行直线标识,进而求得图像的倾斜角度,来实现QR码图像的倾斜校正;后续又针对Hough变换检测直线方法的计算量过大、耗时等不足,提出了通过寻找位置探测图形来定位QR码的三个角点坐标来实现图像的快速定位与校正;对于发生几何形变的QR码图像,提出了一种采用扫描边缘点拟合直线的方法选取畸变校正控制点,利用控制点在几何平面的投影转换进行初步的QR码畸变校正,再通过双线性插值算法进行灰度级的插补,实现了畸变图像的校正。最后,基于MATLAB软件平台将QR码的图像处理与校正算法及编解码等功能整合到同一系统中完成了GUI界面设计,使得整个处理过程可视化,从而更加形象直观地感受到各种算法的实现效果。
[Abstract]:As an emerging electronic payment medium, the two-dimension code has been widely paid attention in various industries, and some developed cities at home and abroad have tried to introduce the two-dimension code payment technology into the subway ticket system, and the two-dimension code recognition technology also becomes the focus in the field of automatic identification research. Since the development of QR code, many advantages of QR code, such as reading speed, storage capacity and error correction ability, are more and more recognized, but it is difficult to introduce some noise, such as uneven image illumination, in the process of QR code image acquisition. The problem of tilting or geometric transformation may bring some degree of error to the decoding of the qr code. Therefore, we should make appropriate image restoration correction for such problems. In this paper, the algorithm principle and basic process of QR code recognition are briefly described, and the image processing algorithms involved are compared and analyzed. The main contents of this paper are as follows: First, the structure, coding principle and basic flow of QR code are discussed in detail, and the recognition process of QR code is divided into two parts: image processing and decoding. Then the process of QR code image processing is described and simulated, and the experimental results are analyzed and compared. At the same time, for the analysis of the damage type of the QR code in a complex environment, a binary improvement algorithm for adaptive brightness is proposed on the basis of image processing, and the problem of the correction and recovery of the QR code image under the non-uniform illumination is solved. In this paper, an adaptive median filter is used to improve the traditional Canny operator edge detection method, and then the edge detection method is combined with the Hough transform to carry out linear identification, and then the inclination angle of the image is obtained, so that the inclination correction of the QR code image is realized; in that follow, the calculation amount of the linear method for Hough transform detection is too large, the time consumption and the like are insufficient, the three corner point coordinates of the QR code are positioned by looking for a position detection graph to realize the rapid positioning and correction of the image, and the QR code image with the geometric shape change is A method for fitting straight lines by scanning edge points is proposed to select the distortion correction control points, and the initial QR code distortion correction is carried out by the projection conversion of the control point in the geometric plane, and the interpolation of the gray level is carried out by the bilinear interpolation algorithm, so that the correction of the distorted image is realized. Finally, based on the MATLAB software platform, the image processing and correction algorithm of the QR code and the functions of encoding and decoding are integrated into the same system to complete the GUI interface design, so that the whole process is visualized, so that the realization effect of the various algorithms is more visually and intuitively felt.
【学位授予单位】：天津理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.4

【相似文献】