基于小波包变换和人类视觉系统的数字水印算法研究
发布时间：2018-11-17 15:38
【摘要】：当前,随着信息技术的不断发展和计算机网络的广泛使用,各种各样的文本、音顷、视频等多媒体产品已经逐渐走进了千家万户,但这些产品的知识权保护等问题受到人们的广泛关注。因此,如何更好的保护创造者的合法权益,已经成为近年来的一个重要研究热点。由于数字水印技术在保证图像信息完整性的同时,很好地解决了数字产品的信息安全等问题,成为近年来信息安全、知识产权保护等问题研究的重要技术热点。 本文在对比国内外现有的图像水印算法基础之上,通过总结分析小波分析的理论基础,根据小波包系数分布特性和人类视觉特性相结合的特征,采用了一种基于小波包变换和人类视觉系统的数字图像水印算法,对数字图像水印进行处理。该算法首先是将水印图像利用Arnold混沌映射进行置乱,然后对宿主图像进行小波包分解,在低频子带处通过改变两个小波子带相应系数的大小关系来嵌入水印信息,再通过控制嵌入的强度,使得水印具有了较好的不可见性和鲁棒性。 本文介绍了数字图像水印技术的研究背景以及现状,阐述了数字水印的基本概念、分类与特性、常见的攻击方法和性能评估指标等内容,论述了小波和小波包变换理论基础,分析研究了数字图像水印的生成、嵌入、攻击、检测与提取等过程。最后,对含水印图像进行加噪、压缩、旋转和滤波等方式攻击性实验,实验表明：该算法水印信息嵌入量大,有较好的鲁棒性和不可见性等优点。
[Abstract]:At present, with the continuous development of information technology and the extensive use of computer networks, various multimedia products, such as text, audio, video and so on, have gradually entered into thousands of households. However, the protection of the right to knowledge of these products and other issues are widely concerned. Therefore, how to better protect the legitimate rights and interests of creators has become an important research hotspot in recent years. Digital watermarking technology not only ensures the integrity of image information, but also solves the problem of information security of digital products. It has become an important technology hotspot in the research of information security and intellectual property protection in recent years. On the basis of comparing the existing image watermarking algorithms at home and abroad, this paper summarizes and analyzes the theoretical basis of wavelet analysis, according to the characteristics of wavelet packet coefficient distribution and human visual characteristics. A digital image watermarking algorithm based on wavelet packet transform and human vision system is proposed in this paper. Firstly, the watermark image is scrambled by Arnold chaotic map, then the host image is decomposed by wavelet packet, and the watermark information is embedded at the low frequency subband by changing the relationship between the corresponding coefficients of the two wavelet subbands. By controlling the embedding strength, the watermark has better invisibility and robustness. This paper introduces the research background and present situation of digital image watermarking technology, expounds the basic concept, classification and characteristics of digital watermarking, common attack methods and performance evaluation indexes, and discusses the theoretical basis of wavelet and wavelet packet transform. The processes of digital image watermarking generation, embedding, attack, detection and extraction are analyzed and studied. Finally, the aggressive experiments of adding noise, compressing, rotating and filtering are carried out on the watermark images. The experiments show that the algorithm has the advantages of large amount of watermark information embedding, good robustness and invisibility.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP309.7
[Abstract]:At present, with the continuous development of information technology and the extensive use of computer networks, various multimedia products, such as text, audio, video and so on, have gradually entered into thousands of households. However, the protection of the right to knowledge of these products and other issues are widely concerned. Therefore, how to better protect the legitimate rights and interests of creators has become an important research hotspot in recent years. Digital watermarking technology not only ensures the integrity of image information, but also solves the problem of information security of digital products. It has become an important technology hotspot in the research of information security and intellectual property protection in recent years. On the basis of comparing the existing image watermarking algorithms at home and abroad, this paper summarizes and analyzes the theoretical basis of wavelet analysis, according to the characteristics of wavelet packet coefficient distribution and human visual characteristics. A digital image watermarking algorithm based on wavelet packet transform and human vision system is proposed in this paper. Firstly, the watermark image is scrambled by Arnold chaotic map, then the host image is decomposed by wavelet packet, and the watermark information is embedded at the low frequency subband by changing the relationship between the corresponding coefficients of the two wavelet subbands. By controlling the embedding strength, the watermark has better invisibility and robustness. This paper introduces the research background and present situation of digital image watermarking technology, expounds the basic concept, classification and characteristics of digital watermarking, common attack methods and performance evaluation indexes, and discusses the theoretical basis of wavelet and wavelet packet transform. The processes of digital image watermarking generation, embedding, attack, detection and extraction are analyzed and studied. Finally, the aggressive experiments of adding noise, compressing, rotating and filtering are carried out on the watermark images. The experiments show that the algorithm has the advantages of large amount of watermark information embedding, good robustness and invisibility.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP309.7
【参考文献】
相关期刊论文 前8条
1 邹建成;基于原根的数字图像置乱技术[J];北方工业大学学报;2001年03期
2 闫伟齐,邹建成,齐东旭;一种基于DES的数字图像置乱新方法[J];北方工业大学学报;2002年01期
3 汪孔桥;数字图像的质量评价[J];测控技术;2000年05期
4 张军,彭宏;基于均值模运算的多小波域水印技术[J];计算机工程与应用;2003年14期
5 尹康康,石教英,潘志庚;一种鲁棒性好的图像水印算法[J];软件学报;2001年05期
6 罗守山,刘s，
本文编号：2338329
本文链接：https://www.wllwen.com/falvlunwen/zhishichanquanfa/2338329.html
