Research On Digital Watermarking Algorithm Based On The Fractional Fourier Transfrom

With the rapid development of IT industry, the spread, copy and storage of digital products becomes very convenient. At the same time, information security problems of the digital products are in urgent needs to solve. Based on hyper chaos theory, fractional Fourier transform and the singular value decomposition theory, the digital watermarking with grey images is studied in this thesis which the main content are as follows:(1) A hyperchaos and fractional Fourier transform based digital watermarking algorithm is proposed. Firstly, we obtain four chaotic sequences generated by hyper chaos system. Then using these four chaotic series, we apply scrambling transform in both position and numerical value to digital watermarking image in order to get the scrambling image. The original carrier image is divided into blocks where the scrambling watermarking image should be put into the appropriate location after the fractional Fourier transform. Analyzing the results from the robustness and invisibility, the experiments show that this method for watermark has well invisibility as well as strong robustness.(2) We also put forward the research and implementation of blind watermarking algorithm based on singular value to quantify the fractional Fourier transform:First of all, the basic theory and characteristics of the singular value are introduced. During the process of the research for this algorithm, we make use of the four chaotic series again to do scrambling transform to the digital watermarking image. Then we divide the original carrier image into blocks, deal with each small block with different fractional Fourier transform order, extract the magnitude of matrix and do the singular value decomposition in order to extract various small singular value matrix. Next we focus on one point in the matrix, quantify it and then embed the watermark. The extraction process of this algorithm does not require the original image, and this algorithm has strong robustness and good visibility. What’s more, this process is simple and has higher practical value.