Digital Image Watermarking Based On Wavelet Transform

With the maturing and diversification of internet and multimedia technology, people can spread, copy, store and process multimedia information, such as images, audio, video, text conveniently. It is a great challenge to multimedia data security, for instance, the unauthorized delivery of video, sound, image and so on. At the same time, it leads the problem of protecting the multimedia information, especially the copyright of digital product. To solve these problems, a new technology, digital watermarking technology which can protect the copyright of digital product and maintain the information security effectively, has been presented recently. Digital watermarking technology demonstrates the ownership of digital product by embedding watermark in the product. Its design key is the invisibility of embedding information and the robustness of resisting various attacks.The development status quo of digital watermark at home and abroad is presented in this thesis, which introduces the basic characteristics, classification and model of digital watermarking technology, analyses the typical watermarking algorithm, and concludes the attack method and certification and evaluation standard. Secondly, the basic principles of continuous wavelet transform and discrete wavelet transform, multi-resolution analysis and Mallat algorithm of discrete wavelet transform is introduced in the condition of one and two dimensional tensor product. It puts forward the construction of compactly supported biorthogonal wavelet and biorthogonal wavelet bases which is used in this thesis, talks about the characteristic and advantage of wavelet decomposition and reconstruction and wavelet domain watermarking algorithm as well. Then a digital watermarking algorithm based on gray image of discrete wavelet transform is proposed, which uses the multi-resolution characteristics of wavelet transformation to decompose the image three layers by compactly supported biorthogonal 9-7 wavelet filter. At the same time, takes advantage of human visual system, error control coding, Arnold transform, pseudo noise sequence of watermarking image and exclusive-OR encryption arithmetic to make the image scrambling encryption, and embed watermarks in y component of the image. Finally, attacking experiments are realized to validate the invisibility and robustness of this method. The watermark attacking experiments include linear and nonlinear filtering, shearing, noise addition, contrast enhancement, and transform-based image compression. In addition, the error rate of each wavelet subband is calculated after various attacks. The results turns out that the method proposed in this thesis is imperceptibility and robustness.