Research On Digital Image Watermarking Technology

With the rapid development of digital technology and the increasing popularity ofthe Internet, digital watermarking as an effective method to protect the copyright ofdigital products has drawn wide attention. In recent years, the wavelet technologydevelopment tend to mature and its application in digital watermark is more and morevalued, blind watermarking algorithm based on wavelet is one of the hot spots in thepresent research. But the most blind-watermark algorithms based on the wavelet are notso good at invisibility and robustness, so this thesis studied more in balancedmulti-wavelet, and proposed a blind watermarking algorithm based on combination ofwavelet and balanced multi-wavelet.This thesis studied the basic theory, features, and technology framework of digitalwatermarking technology, and summarized the commonly used digital watermarkingassessment method. In this thesis, the improved arnold transform algorithm and Henonchaos mapping system are combined in the preprocessing part of the watermark,realized the watermark image position scrambling and pixels scrambling, so as toimprove the robustness and safety of watermark. Based on this, this article also studiedwavelet and balanced multi-wavelet transform theory, then put forward a kind of blindwatermarking based on wavelet combining with balanced multi-wavelet transform. Thealgorithm which analyzes the coefficient of the embeded area implements the blindextraction of the watermark and simplifies the extraction and test process of watermark.In addition, compared with the relevant similar algorithm, after already somewatermarking algorithm usually because of judge selected according to the experienceof the embedding strength coefficient and great subjectivity, and the conclusion is basedon a large amount of repeated try take based on the low efficiency. Therefore this thesisput forward a kind of watermarking embedding strength coefficient of optimizationscheme, the improved genetic algorithm is used for the quantitative analysis, to searchfor appropriate global optimal embedding strength coefficient, to optimize for theinvisibility and robustness of the watermark.Experimental results show that in premise that the information of watermark isinvisible and the algorithm has good robustness for the common JPEG compression,filtering, adding noise and other attacks. It will study the image watermark algorithm isextended to multimedia in the future.