【作者】 吴松；

【导师】 赵铁军；

【作者基本信息】 哈尔滨工业大学 ， 计算机技术， 2012， 硕士

【摘要】 QR码具有信息容量大、可靠性强、识别速度快、识别范围广、编码种类多等优点，在许多行业中得到了广泛的使用。隐形QR码将隐形印刷技术与QR条码相结合，使得QR码拥有了隐形印刷优秀的防伪能力。但由于印刷、识读方式不同，与非隐形QR码相比，隐形QR码图像复杂，将QR码与背景分离更困难。如何从隐形QR码图像中分离背景和QR码信息，识别出深浅模块；寻找位置探测图形，根据校正图形中心对QR码数据取样是本文研究的重点。论文通过理论分析和实验结合，主要研究了QR码的图像识别技术，主要内容如下：(1)针对所采集的隐形QR码图像在紫光照射下颜色失真和拍摄时光照不均等因素而产生的图像背景与QR码模块难以分离的困难。论文首先根据隐形QR码图像R通道QR码信息突出的情形，对QR码进行了灰值计算。然后采用了自适应图像二值化方法，识别出QR码的深浅模块。(2)针对隐形QR码图像背景复杂，深浅模块识别有部分错误，使QR码图像难以定位的困难。根据隐形QR码的寻象图形特点，修改了QR码规范中图像定位的方法，提高了寻找QR码位置探测图形的准确率。(3)针对采集QR码图像时因拍摄角度产生的QR码图像畸变问题，参照QR码规范中的图像采样算法提出了有效解决畸变图形采样的方法。该方法首先根据位置探测图形逐次寻找校正图像中心点，然后根据校正图形的位置和寻象图形构成的取样网格对QR图形进行取样。(4)针对隐形QR码图像质量差，为提高识别率，选择纠错级别H的纠错码字对数据进行纠错。实现了RS纠错码在GF（28）域上的映射表和逆映射表。(5)为进一步加强隐形QR码的防伪能力，选择数据加密算法对输入数据进行加密。更多还原

【Abstract】 Because of its advantage of large information capacity, high reliability, fastrecognition speed, wide recognition range, various encoding categories, quickresponse code (QR code) has been widely used in many fields. Invisible QR codemakes QR code inherits the brilliant security capabilities from invisible printing.Invisible QR code images are more difficult to separate the QR code and thebackground. This paper puts emphasis on how to separate the QR code informationand the background from the invisible QR code, how to read the data from the QRcode and improve the accuracy of invisible QR code recognition. Throughtheoretical analysis and experiments, this paper mainly studys the invisible QR coderecognition technology, the main content of this paper is as follows:1It is difficult to separate image background from the QR code module for thecolor distortion and the uneven illumination when shooting the invisible QR. Tosolve this problem, this paper firstly calculates the grayscale value of the QR codeaccording to the situation of the QR code information in invisible QR code image’sR channel is outstanding. After that, using the adaptive grayscale image binarizationmethod, this paper separates the QR code and the background and recognizes thedifferent depth of modules.2Because the background of image is complex and there are some errors whenrecognizing the different depth of modules, it is hard to locate the QR code image.According to characteristics of the invisible QR code image, this paper modifies theimage location method in QR code rules to improve the accuracy of finding QRcode location detecting images.3The sampling images can be distorted when collecting the QR code images.With reference to the image sampling method in QR code rules, this paper proposesa method to solve the sampling image distortion. In this method, the first step is tofind the correction graphics center, and the next step is to sample the QR codeimages using the area network established by the location of the correctiongraphics.4To improve the accuracy of recognizing the invisible QR code images, thispaper chooses a error correction code word with correction level H to correct thedata. This paper also implements the RS error correction code mapping table andinverse mapping table in the primitive elements of GF(28).5To further enhance the security capabilities of the invisible QR code, thispaper chooses the AES method to encrypt the input data.更多还原