Research On The Digital Watermarking Videoalgorithm Based On Wavelets Domin And Image Fusion

With the increasing popularity of computer networks and the rapid development of multimedia technology, a number of digital productions can conveniently transmit in the Internet ,but at the same time piracy and copyright issues are becoming increasingly serious, so information security and copyright of digital production have attracted attention highly, new digital watermarking techniques is desiderate to resolve the problem. Because the digital products have widespread redundant and randomness data,the information such as digital,serial number, text,images and other signs are embedded in the products itself, so is digital watermarking technology. Now digital watermarking technology is more and more important,which will help a copyright protection,secret communications,and the authenticity of data files,and product identification and so on.Video watermarking technology is a hot and difficult point in current digital watermarking technology study. This article mainly aims at the video copyright protection question, carries on the the research of video watermark technology.This thesis related these points and the main contribution includes the following parts:(1) For this paper,some important topics about digital watermarking are discussed,and some basic knowledge of the relevant video watermarking is presented, including application background, present research situation, technical classification and performance requirements(2)The selection of video watermarking embedded field. According to the relevance of video inter-frame and inner-frame,an motion estimated method is proposed based on wavelet. This method performs wavelet decomposition for the video frame. It determines each picture background by the low-frequency coefficient and definite picture details by the high -frequency coefficient. Compared the difference and threshold of wavelet coefficient between the neighboring frame, the key frame is analyzed both for different and identical scenes.(3) In order to make Haar wavelet function to be more accurate, and multi-resolution analysis to reflect the video characteristic well, this paper improves Haar wavelet so that it has orthogonality, symmetry and frequency range smooth.(4) In order to improve the security of the watermark, the use of a method of Arnold to scramble watermark, while the scrambling watermark is divided into five pieces.The five pieces of watermarks are embedded into different scene, a certain watermark is embedded into Y channel of each key frame in the same scene, which enhances the robustness of the watermark.(5).Applying human visual system and image fusion techniques to video watermark is the major novelty of this paper. In embedding watermark process, video frame images and scrambled watermark images are wavelet transformed simultaneously, and then the watermark was added to the third level of original video frames images. The high and low frequency coefficients of wavelet transformed watermark images and video frame images are also merged, insertint the original watermark to the video compression sequence ,so the watermark is more robust.Experiments show that,the algorithm of this thesis has better robust to most of attack and greatly invisibility, and achieves the binding detection. The function is implemented by Matlab and the algorithm has more universality.