Research Of Digital Watermarking Based On Chaotic System

With the rapid development of Internet technology, the interconnection of network becomes more and more simple and increasingly popular. The infinite connectivity and the sustainable development of digitalization of technology information media make it easier for person to acquire all kinds of media data from Internet, such as text, picture, audio, video and so on. Because of the perfect reproducibility and non-destructive of digital media data, an incidental side-effect was that a large number of pirating problems appeared. Therefore, in the infinite interconnected environment, how to effectively protect the copyright of digital media data becomes very important and an urgent problem in practice. At the same time, with the increasing trend of the popularization of information media and rocketing development of Internet, digital media data is exposed to the public easily. The original digital media data do not have enough safety protection, people can modify and change it freely, especially for some privacy data, thus it had some negative influences on the original digital media data. Consequently, how to effectively protect and authenticate the content of digital media data and its completeness become another very important and an urgent problem in practice, together with copyright protection become two very important questions which urgently wait to be solved at present. Digital watermarking is a potential method possible to solve the above problems. So it has attracted the attention of government, enterprise field and international academic community.This thesis mainly explores image watermarking algorithms on the basis of wavelet-domain and wavelet packet domain, and chaos system is used for encrypting the watermarking, coupled with flexible usage of genetic algorithm, the main contents include the following:(1)The subject background and the significance of the article are introduced and the current state and trends of development of digital watermarking is described. It overviews various methods of image watermarking algorithms existed at present, then later on point out the existent problems and development tendency of digital watermarking.(2)Give a sketch of the definition and the basic framework of digital watermarking、three elements of digital watermarking、performance analysis and classification of digital watermarking、some of the common image processing attacks and applications of digital watermarking.(3)Looking back the history of chaos theory and giving a brief illustration of the definition of chaos、the characteristics of the chaos and chaos system are applied in the digital watermarking,it is emphatically introduced several typical chaos system.(4)An algorithm of digital image watermarking based on block energy analysis of the wavelet transform coefficients and chaos system was put forward. Arnold transform is used together with Lorenz chaotic systems to encrypt watermark in order to enhance the security. Make full use of chaotic matrix secondly to establish a new relationship among the chosen wavelet coefficients. According to this new relationship embed the watermark information into the largest wavelet coefficients whose block energy bigger in order to improve the robustness of the algorithm. No original image is required for watermarking extraction by taking advantage of the same relationship. The result of experiments shows that this algorithm has satisfactory performance with low computation complexity, transparency and robustness, but this algorithm is disadvantageous in the embedding capacity is small.(5)Combining genetic algorithm and Logistic chaos sequence, wavelet tree are optimized by genetic algorithms, then the best wavelet base is found out, as a result, the transparency and robustness of watermark is improved. Through the highly innovative use of Logistic chaos sequence, enhancing the complexity and randomness of chaos sequence, so it greatly improves the security. Try to balance the transparency and robustness, the embedding positions are chosen at random. No original image is required for watermarking extraction by taking advantage of the relationship among the chosen wavelet coefficients. The simulation results show that the transparency and robustness are perfect, besides, the embedding capacity is large. The only fly in the ointment was that the computation complexity of this algorithm is high.