Research Of Multiple Watermarking Algorithm

With the rapid development of multimedia technology and computer network technology, and the growing phenomenon of piracy, the copyright protection of digital products is particularly important. Digital watermarking technology, as a major branch of information hiding technique, is an important means of copyright protection, and attracts more and more attention of overwhelming majority of scientific and technological researchers and users, and research results are also emerging in endlessly. They still can not solve the problems of multiple copyrights and copyright certification of digital products at different stages of the release, sale, use and so on. So multiple watermarking technology emerges as the times require. Multiple watermarking algorithm has more requirements than a single algorithm in the aspects of watermark generated and algorithms embedded. For example, watermark must be generated simply, and the large number of watermarks should have non-interference in the process of watermark embedded and tested, and the transparency and robustness of the watermarks are more complicated. One of the most important issues is how to ensure that multiple watermarks embedded have no effect on each other.After analyzing the present research status of digital watermarking at home and abroad, an improved single-value watermarking algorithm based on the LSB was proposed. First, the binary watermark image was changed the same size as the host image by over sampling. Then Arnold transform was employed for the host image and binary watermark image changed. The bit embedded, which is one of the neighboring bits, is decided by the chaotic sequence. Finally, Arnold inverse transform was used to obtain the image which was embedded watermark.A multiple blind watermarking algorithm was put forward, which bases on the composition of color images and embeds watermarks in the spatial domain and wavelet domain of the gray images decomposed. First, the host color image was decomposed into three gray-scale images of RGB, then the algorithm in the spatial domain was used in the RG components, which were embedded watermarks in the low bits, and the algorithm in wavelet domain was used in the B component, which was embedded watermark in the low frequency coefficients of 4-level wavelet decomposition.With the application of discrete cosine transform, spread spectrum and BCH coding and decoding techniques, a multiple blind watermarking algorithm in the DWT and DCT was presented, which bases on a spread spectrum. It based on the previous algorithm of multiple watermarking in the color image, and used the spatial domain in the R component, used the DWT in the G component, used the DCT in the B component after blocking. Because of the discrete cosine transform, spread spectrum and coding and decoding techniques, the security and robustness of the watermarks were enhanced.The attacks including salt and pepper noise, Gaussian noise, Poisson Noise, multiplicative noise, linear spatial filtering, image shearing, JPEG compression and so on were done to images embedded a single watermark or multiple watermarks. The experiment results show that the scheme can not only be able to guarantee the security of the watermark itself, but also has good robustness and visual concealment. This method does not require the original image or original watermarks when extracting watermarks, and successfully tackles the interaction of watermarks embedded in one image.