基于混沌加密与奇异值分解的图像水印算法研究

【作者】 李迎江；

【导师】 冯乔生；

【作者基本信息】 云南师范大学 ， 计算机软件与理论， 2007， 硕士

【摘要】 近年来,多媒体技术和宽带网络技术的不断进步,促进了数字媒体应用的迅猛发展,各种形式的多媒体作品,如图像、音乐和电影等迅速地以网络的途径进行传播,这一方面便利了人们相互间的信息交流,但另一方面也给未授权的第三方访问并操作多媒体作品提供了渠道,严重地侵害了作者的权益。保护多媒体产品的安全成为当前的研究热点之一数字水印技术作为信息隐藏技术的一个重要分支,是一门新兴的信息安全技术,它是通过在数字产品中嵌入版权信息——水印来证实该作品的所有权,被嵌入的信息有的是可以直观看到的,有的是不可见或不可察的,但是通过对水印载体的检测仍然可以提取出水印或者判定水印是否存在。目前,数字水印技术己在多媒体作品的版权保护、拷贝追踪以及完整性认证等领域体现出巨大的应用价值。对于不可见的鲁棒的数字水印研究的文献较多,但是到目前为止,大多数这种水印对几何攻击的鲁棒性都较差,对此问题本文进行了研究,主要工作如下：1.改进了一种基于混沌序列的图像加密方法,此方法利用两个混沌序列分别对图像每个像素的横纵坐标同时进行置乱加密,混沌序列的两个初始值作为密钥,这种方法进一步增强了加密算法的抗破译能力,使得经此方法加密后的水印图像更加安全。2.对一种已有的基于奇异值分解的水印算法进行研究,并发现这种水印算法存在问题,通过实验证明,这种水印嵌入算法即使在没有嵌入水印的图像中也能够提取出水印,而且提取出的水印还比较清晰。并在理论上证明了此方法的问题的所在。3.本文提出一种新的水印嵌入算法,此算法是先将原始图像分块,选择处于平滑区域的块进行SVD分解,再将水印经混沌加密和一维化处理后嵌入到每一块的奇异值中,经逆SVD分解得到嵌入水印的图像。在提取过程中,先对待测图像进行几何变换估计并还原,然后对水印进行提取,最后对此算法进行大量实验,并与已有算法进行比较,表明此算法抗攻击能力更强。4,对于在水印提取过程中用到的已有图像几何变换估计方法进行分析,发现其方法中的旋转变换估计平移变换估计的实际应用价值不大,对此本文提出一种新的旋转变换和平移变换估计的方法,算法更简单,且更有实用价值。随着互联网的迅速发展,数字多媒体产品(图像、视频、音频、文档等)比以往更加容易获取、复制和传播,因此保护这些数字作品的版权就成为一个重要的课题。数字水印就是保护数字产品版权的一种有效技术,它是通过在数字产品中嵌入版权信息——水印来证实该作品的所有权,被嵌入的信息通常是不可见或不可察的,但是通过对水印载体的检测仍然可以提取出水印或者判定水印是否存在。该技术自20世纪90年代初提出至今,已经引起了许多专家学者和商业团体的广泛关注,成为国内外研究的热门课题之一,研究成果也层出不穷,但数字水印技术还远未成熟,许多问题亟待解决。本文是对图像数字水印进行综述性研究和阐述,首先第一章回顾了数字水印的研究背景及研究意义,介绍了数字水印的应用领域。接下来第二章对数字水印的技术,例如：数字水印的概念、特点、基本框架、常用水印算法进行了详细的阐述。第三章对数字水印的分类进行介绍,水印根据载体数据、实现算法、视觉效果、特性等分成多种类型,不同的类型有不同的应用场合。并描述了对含有水印的媒体信息在传播过程中可能受到各种有意或者无意的攻击,以及对于各种攻击应有的对策。第四章着重讲述了数字水印的置乱加密技术,为了增强数字水印的安全性,一般水印算法都会在嵌入前对水印信息进行加密,这样即使水印被提取,也不会暴露水印信息,而且会加强水印的抗攻击能力。第五章对当前常用的水印算法进行分析和评价,基于空域的水印算法是将水印信息直接嵌入到图像像素位中,算法简单,但抗攻击能力差,近年已很少出现；基于变换域的水印算法是将水印信息嵌入到图像的变换域中,使嵌入的水印更不易察觉,且抗攻击能力更强,特别是基于DCT,DWT等变换的水印算法成为当今比较热门的算法。还有一些结合多种变换域进行水印嵌入的算法也取得了很好的效果。更多还原

【Abstract】 In recent years, the constant progress of multimedia technology and broadband network technology, has promoted the swift and violent development that the digital media used. The multimedia works of different forms, for instance image, music and film,etc. spread with the way of the network rapidly, on one hand this has facilitated people’s mutual information interchange, but even visit and operate the multimedia work and offer the channel for the authorized third party on the other hand, for example anyone can get others’ primitive works easily through the network, protect the security of multimedia products has becoming one of the focus of researchAs an important branch of information hiding technology, digital watermark technology is a new developing information safe technology, it is through imbedding copyright information in the digital products-watermark to verify the ownership of works, some information that imbedded is can be seen, some can’t be seen or examined, but can still draw out the watermark or judge whether the watermark exists through the measuring of the watermark carrier. At present, digital watermark technology embody enormous using value in such as protect copyright protection, copy tracing and integrality authentication, etc.There are more documents studied to the robust, unseen digital watermark, but up till now, its robustness is relatively bad to geometry attacked in great majority of this kind of watermark, this text has carried on research in these question, the main work is as follows:1. Put forward a kind of new image encryption method based on Chaos array, this method utilizes two Chaos arrays to put encrypting to the horizontal and vertical ordinate of each pixel of image at the same time separately, two pieces of initial value of the Chaos array as the key, this kind of method has strengthened resisting the ability of decoding of the algorithm of encrypting further, make the watermark image safer after encrypted by this method.2. Carry on research to a kind of existing watermark algorithm based on SVD, And find that the problem exists in this kind of watermark algorithm, prove through the experiment, this kind of watermark embed the algorithms can draw out the watermark even in not imbedding the image of the watermark, and the watermark that draw out is also relatively clear. Meanwhile, proved the question of this method in theory.3. This paper proposes a kind of new watermark imbeds algorithms, this algorithm divides the primitive image first, select the blocks in smooth piece of area carry on SVD resolve, embed the watermark by Chaos encrypted and one dimension changing into each block’s SVD value, got the image which imbeds the watermark against SVD. In the course of drawing,estimate the geometry transform the image and revert first, then draw the watermark, carry on a large number of experiments to this algorithm finally, and compared with exist algorithms, prove that this algorithm has a stronger ability in resisting attack.4. Analyses the existing method of image geometry estimated during the course of recover watermark, find that the rotation estimates method is not very useful in actual using,in this paper, put forward a kind of new rotation estimate method, and the algorithm is simpler, and more practical value. With the rapid development of Internet, the digital multimedia products (picture, video, audio, document, etc) are easier to obtain, duplicate and spread than in the past, so, protect the copyright of these digital works and an important subject. The digital watermark is a kind of effective technology of protecting the digital products copyright, it is through imbedding copyright information in the digital products-the watermark to verifies the ownership of this works, the imbedded information usually can’t be seen or can’t be examined, but can still draw out the watermark or judge whether the watermark exists through the measuring of t he watermark carrier. This technology was proposed since the beginning of 1990s, have already caused the extensive concerns of a lot of experts and scholars and commercial organization, becoming one of the hot subjects studied both at home and abroad, the research results emerge in an endless stream too, but the digital watermark technology is still far from ripe, many problems are demanding prompt solution.This paper carries on research and explains to the digital watermark of the picture, chapter one has reviewed the research background of the digital watermark and research meaning at first, introduced the application of the digital watermark. And then chapter two, about the technology of the digital watermark, for example:The concept, characteristic, basic frame, common used watermark algorithms carried on detailed exposition. Chapter three make an introduction to the classification of the digital watermark, the watermark is divided into many kinds according to carrier data, realizing algorithm, visual effect, characteristic, etc., different types have different application occasions. And described that may be attacked media’s information with watermark by various kinds of ones that intend or have no intention in the course of propagating, and the countermeasure due to various kinds of attacks. Chapter four tell us the encryption techniques is put in digital watermark, in order to strengthen the securities of the digital watermark, generally, we will encrypt watermark information before imbedding in the watermark algorithms, even if the watermark is drawn, it will not expose watermark information either, and will strengthen the ability of resist attacking.Chapter five analysis and appraise about the present commonly used watermark algorithm, the watermark algorithm based on airspace imbeds watermark information to the picture’s pixels directly, the algorithm is simple, but it is bad to resist attacking, already seldom appeared in recent years; The ones that imbed watermark information to the picture’s transform area, make the watermark imbedded more difficult to perceive, and it is stronger to resist attacking, especially the DCT, DWT, nowadays becomes hotter algorithm. And some combined many kinds of transform area algorithms got good result.更多还原