Design And Implement Of Digital Image Watermark Algorithm Based On Chaos Theory And Wavelet Transform

As an important means of copyright protection, digital image watermarking technology is the hotspot in the field of image protection technology research in recent years, and wavelet analysis is a frontier theory highly concerned by the current international scientific and academic community. Facts have proved that, among numerous digital watermarking algorithms, the wavelet transformation-based algorithm has more superior characteristic than others, and could effectively protect the copyright of digital products.To resolve the issue of uncertainty, poor safety and easy-decoding existed in the current digital watermarking method, the paper puts forward a wavelet transformation-based algorithm: conduct three-level wavelet decomposition on images, adopt different methods at low and intermediate frequencies to embed binary image watermark, and before embedding watermark, use Arnold transformation to conduct scrambling processing, therefore realizing data confidentiality and enhancing data security. The following are the main points:(1) The paper analyzes and researches the basic concepts, features and development situation at home and abroad of digital watermarking technology, chaos theory and wavelet transformation, as an algorithm the solely wavelet transformation-based watermarking has some shortcomings in robustness and transparency.(2) To resolve the issues above, the paper puts forward a chaos-based wavelet-domain digital watermarking algorithm. Use a binary image with special meaning as watermark, preprocess watermark before embedding it to scramble original watermark into messy information, and then embed watermark, therefore strengthening the safety and invisibility of watermark. The algorithm in the paper chooses binary image'Hunan University'as watermark, encrypt it with Arnold chaotic sequences, and embed the scrambled watermark in wavelet coefficients of different directions according to the characteristic of wavelet transformation and in combination with the Human visual system(HVS).(3) To test the performance of the watermarking algorithm, the paper conducts various attack tests such as still image compression (JPEG), noise adding, Gaussian filter, average filter, image size changing, image rotation, etc on the watermark image by MATLAB. Experimental results show that the watermarking algorithm raised by the paper has relatively strong robustness and transparency.(4) To embed more watermarks without affecting original image quality and in combination with the HVS, the paper puts forwards chaos theory and wavelet transformation-based watermark embedding and extraction algorithm. Experiment results have shown that the improved watermarking algorithm could well balance the invisibility and robustness of watermark and have relatively strong robustness against noise, compression and cut attacks. It is a feasible algorithm, having contributed positively to the current wavelet-domain digital watermarking algorithm.