The Research Of Digital Watermarking Technology Based On Chaos Theory And Discrete Wavelet Transform

With the rapid development of multimedia information technology and Internet, the problem of the copyright protection has become more and more serious. Digital watermarking, as an effective method to protect the copyright, has been receiving extensive attentions.This paper focused on the excellent properties of the chaotic sequence and DWT in digital watermarking applications. The following are the main points:First of all, this paper have summarized the essential features, principles and domestic and international research situations of digital watermarking, chaos theory and wavelet transform, and have analyzed the digital watermarking algorithms and the methods of watermarking attack in details. On this basis, the purpose of the paper was proposed.Secondly, a watermarking encryption algorithm based on chaos theory was proposed. This algorithm, in which a binary image with practical meaning was selected as the watermark, and encrypted by the chaotic sequence generated from composite nonlinear discrete chaotic dynamical systems. The initial values that generated the chaotic sequence can be taken as the keys. Only obtain all of the keys that you can extract the correct watermark. We also did the simulation to prove the algorithm's feasibility.Furthermore, we proposed a watermarking algorithm based on chaos and DWT. In this algorithm, we took the meaningful binary image "SDUST" as the Watermark. Before embedded it into the carrier image, we used the LOGISTIC chaotic sequence to encrypt it. Combined the HVS (human vision system) models and according to the characters of wavelet transform, we adopted multi-resolution analysis technology and took the way of corresponding multi-resolution embedding after the watermark and the original image were resoluted with wavelet transform. For enhancing the robustness of the watermark algorithm, we embedded the watermark repeatedly to the large coefficients of the low-frequency. Thus the algorithm's invisibility and robustness request was advanced.At last, in order to further evaluate the algorithm, MATLAB simulation experiment was done include the attack of JPEG compression, common noise, Gaussian filtering, Median filtering, Average filtering, image scaling, image rotating and so on. And the results proved the algorithm's robustness.