| As rapid development of the information technology and computer network technology, more and more images, videos and other digital multimedia products are created and spread in the network. It also makes the field of digital multimedia havea need of a safe and effective copyright protection technology. As an important branch of information hiding area, digital watermarking technology is created to provide a new way of thinking for copyright protection for multimedia provides. Digital watermarking technology will embed watermark containing specific information digital media, and extraction and identification the watermark to confirm the copyright information the time needed.The research object of the paper makes is the popular MPEG-2 video, and introduces the basic framework of digital watermarking technology, and a detailed analysis of video watermarking technology. In this paper, decoding and extraction of I frame in MPEG-2 and feature extraction of image points were the two key research areas, and ultimately on this basis, a watermarking MPEG-2 video frame based on characteristics of image points was proposed which will be embed watermark on the corner in I frame. Finally, the simulation test program. The main research and innovation are as follows:(1) The watermark should be embedded in I frame, so I frame is extracted watermarking system must be the technical aspects. In this paper, MPEG-2 video compression standard for detailed was analyzed and researched, combined with MPEG-2 video elementary stream structure, this paper design and implement I frame extraction algorithm in domain in the video compression code stream; then crop the MPEG-2 decoding algorithm in the standard documentation, so it apply to the stream to decode the I frame only. After the completion of these two processes to achieve the only MPEG-2 video decoding I frames. Compared to decoding the whole MPEG-2 video and then extract the I frames, the former obviously much more efficient.(2) The watermark bits will be embedded in the corner of I frames, so this paper analyzes the characteristics point of the image, focus on the Harris operator, first implemented a method based on Harris corner detection operator, then improvements it, using the matrix determinant and the trace to calculate the corner detection operator, in order to avoid complicated matrix eigenvalue decomposition, to accommodate a large number of I frames corner detection needs. Video data is usually large, simplify the Harris operator to the system can increase the efficiency of the watermark. (3) According to the proposed video watermarking algorithm based on the MPEG-2 corner of I frames,a complete digital watermarking system was achieved, and the concealment and anti-attack tested. The results show that in the visual range of people, it is difficult to distinguish the change of the watermarked video before and after watermarking; in the absence of any attack and interference by the case, the embedded watermark in the video vector can be completely extracted correctly; to be geometric attacks and the bit rate higher than 1Mbit/S compression attack, the watermark bit can also be successfully extracted and identified. |
PDF Full Text Request