结合QR码的电子发票图像认证水印算法研究

发布时间：2018-12-30 21:52

【摘要】：发票是企业会计记账和纳税申报的原始凭证,随着电子商务的发展,纸质发票与电子商务交易的即时、快速已不相适应。电子发票的推行有助于节能环保、规范税费管理、降低运行成本、提高运营效率,但是电子发票造假行为也日益猖獗,为实现电子发票图像真的实性认证,本文以数字水印技术为基础,设计了一种结合QR码的电子发票图像认证水印算法。论文首先对研究意义与背景进行了介绍,指出根据票面构成特点,将电子发票图像的真实性认证分为QR码区域和非QR码区域两部分,在实现双重认证的同时,提高了认证效率。主要研究内容为以下两个方面：(1)基于QR图像的隐写,改进算法基于两个特点提出：黑白模块比例接近1：1、解码时允许每个模块有一定范围内的偏差。为便于网络传输及存储,选择以二值图像为研究对象；为了保证载密图像能被正确解码,在允许范围内调整模块的边界；为了图像的增大隐写容量,在模块交界处嵌入秘密信息；为了增强算法的安全性,以密钥控制秘密信息的嵌入位置。也即本文算法是通过调整黑白模块的边界的方法,将秘密信息嵌入模块交界处,并以密钥控制嵌入位置。实验结果表明,本文算法在隐写容量、抗缩放攻击方面的性能表现要优于现有算法,但是对打印扫描攻击不具有抵抗性。(2)根据电子发票图像特点,选取GIF格式的发票图像为研究对象。第一重认证：将电子发票信息作为水印嵌入QR图像实现篡改检测。第二重认证：结合发票图像颜色信息有限的特征,将调色板进行重写扩充,通过改变索引矩阵的方式嵌入水印,达到了较好的人眼“不可见”。首先对图像分块分类,对每一图像块生成认证水印并嵌入自身块；仅对每一个重要类图像块以“二值化”的方式生成恢复水印,并基于密钥将其嵌入在两个非重要类图像块内。对于只需要认证发票真实性的场合,可以仅进行第一重认证,如果需要进行定位与恢复,再继续进行第二重认证,可以大大提高认证效率。论文的最后,针对本文的研究成果设计了一个可视化人机互动仿真系统。
[Abstract]:Invoice is the original proof of accounting and tax declaration. With the development of electronic commerce, the paper invoice is not suitable for the transaction of electronic commerce. The implementation of electronic invoice is helpful to save energy and environmental protection, standardize tax and fee management, reduce operation cost and improve operation efficiency. However, electronic invoice fraud is becoming increasingly rampant, in order to realize real authentication of electronic invoice image. Based on digital watermarking technology, a digital invoice image authentication watermarking algorithm combined with QR code is designed in this paper. This paper first introduces the significance and background of the research, and points out that the authenticity authentication of electronic invoice image is divided into two parts: QR code area and non-QR code area according to the characteristics of face structure. The authentication efficiency is improved while double authentication is realized. The main research contents are as follows: (1) based on the steganography of QR images, the improved algorithm is proposed based on two characteristics: black and white module ratio is close to 1: 1, and each module is allowed to have a certain range of deviations when decoding. In order to facilitate network transmission and storage, the binary image is chosen as the research object, and the boundary of the module is adjusted within the allowable range in order to ensure that the carrier image can be decoded correctly. In order to increase the steganography capacity of the image, the secret information is embedded at the junction of the module, and in order to enhance the security of the algorithm, the embedding position of the secret information is controlled by the key. In other words, by adjusting the boundary of the black and white module, the secret information is embedded in the interface of the module, and the embedding position is controlled by the key. Experimental results show that the performance of this algorithm in steganography capacity and resistive scaling attack is better than that of existing algorithms, but it is not resistant to print scan attack. (2) according to the characteristics of electronic invoice image, The invoice image in GIF format is selected as the research object. First authentication: the electronic invoice information is embedded into QR image as watermark to implement tampering detection. Second authentication: combining with the limited color information of invoice image, the palette is rewritten and expanded, and the watermark is embedded by changing the index matrix to achieve a better "invisible" human eye. Firstly, the authentication watermark is generated for each image block and embedded into its own block. Only the watermark is generated "binary" for each important class image block, and embedded in two non-important class image blocks based on the key. For the situation that only needs to authenticate the authenticity of invoice, we can only carry out the first authentication. If we need to locate and restore, and then continue to carry on the second authentication, the efficiency of authentication can be greatly improved. At the end of this paper, a visual man-machine interactive simulation system is designed according to the research results of this paper.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP309.7

【相似文献】