【作者】 张明；

【导师】 付炜；

【作者基本信息】 燕山大学 ， 通信与信息系统， 2009， 硕士

【摘要】 随着网络技术和多媒体技术的迅猛发展,数字水印技术已经成为版权保护的一种有效办法。音频水印是以音频为载体的水印技术,通过在音频载体中嵌入水印信息,实现版权保护、掩蔽通信的目的。目前,国内外数字水印都是基于静态图像的水印嵌入研究,对音频水印的研究极少。本文以二值图像为水印信息,以音频数据为嵌入对象,在分析了数字音频水印的特点和声音信号的理论基础上,从提高水印的不可察觉性和鲁棒性出发,对现有的水印算法进行改进,提出了两种数字音频水印的算法。一是针对现有音频水印算法缺乏有效的同步技术,提出了一种基于离散小波变换的同步水印算法。算法引入了同步信号,利用同步信号定位水印的嵌入位置。为了提高信号的鲁棒性,把水印与同步信号都隐藏在小波域的低频系数中,检测水印时不需要原始的音频。实验表明,该算法具有较强的鲁棒性和不可感知性,并可以抵御各种常见的攻击。二是针对单音频水印算法的局限性,提出了一种基于DCT和DWT混合域的音频水印的算法,首先对原始的音频信号进行离散小波变换,得到音频信号的精细分量和近似分量。然后对这两种系数进行离散余弦变换,最后将两种不同的水印嵌入到两种系数。提取水印信息时,不管音频信号的长度如何,均采用相同的方法进行变换,实验证明,该方法可以实现不丢失音频信号的频率信息,在考虑嵌入的两种不同水印的顺序的原则下,实现嵌入和提取信息的同步。最后结合目前音频水印领域的研究成果,提出了算法存在的问题和局限性,明确了以后工作的研究方向。更多还原

【Abstract】 With the network technology and the rapid development of multimedia technology, digital watermarking technology for copyright protection has become an effective approach. Audio watermarking based on audio watermarking technology for the carrier, through the audio watermark embedding research on audio watermarking few studies. In this paper, binary image for the watermark information to audio data for the embedded object, the analysis of the characteristics of digital audio watermarking and audio signals on the theoretical basis, from the watermark can not improve detection and robustness, the watermark of existing algorithm, two digital audio watermarking algorithms.First, the existing audio watermarking algorithm for the lack of an effective synchronization technology, a discrete wavelet transform-based watermarking algorithm synchronization. Algorithm for the introduction of synchronous signals, the use of synchronous signal the location of embedded watermark positioning. Signal in order to improve the robustness of the watermark and the synchronization signals are hidden in the low-frequency coefficients in wavelet domain, the detection of the watermark without the original audio. The experiments show that the algorithm is robust and can not be perceived, and can withstand a variety of common attacks.Are two single audio watermarking algorithm against the limitations of a hybrid based on the DCT and DWT domain audio watermarking algorithm, the first of the original audio signals, discrete wavelet transform, the audio signal and the approximate weight of the fine component. And these two discrete cosine transform coefficients, and finally two different watermark is embedded into the two factors. When extracting the watermark information, regardless of how the length of the audio signal are used to transform the same way, experiments show that the method can not realize the frequency of the loss of audio signal information, in considering the embedded watermark sequence of two different principle, the implementation embedding and extraction of the synchronization information.Finally, the current research in the field of audio watermarking results, proposed algorithm problems and limitations of clear direction for future research work.更多还原