Research On Robust Watermarking Algorithm Based On Transform Domain

With development of the Internet,people can easily obtain digital products.Lawbreakers make use of the characteristics of digital products that are easy to be copied and tampered to gain benefits illegally,which infringes the rights and interests of copyright owners.Therefore,we need to find effective ways to protect the interests of copyright owners.There are often some redundant information in digital works.According to the characteristics of redundancy,digital watermarking technology adds watermarking information which is not easy to detect but can determine the difference in digital works,and determines the copyright of digital works according to the extracted watermarking information.Digital watermarking technology effectively safeguards the rights and interests of the original authors,and gradually occupies a major position in the field of multimedia information security.It has become a research hotspot of image processing and information hiding technology.Starting from the above background,the related technologies of digital watermarking are studied in depth.The specific research contents are as follows:Firstly,in the process of image preprocessing,a hybrid encryption method using three-dimensional Arnold scrambling and Logistic scrambling is proposed,which improves the security,and three-dimensional Arnold scrambling is more efficient than two-dimensional Arnold scrambling.Secondly,in the Discrete Cosine Transform(DCT)domain,an algorithm for extracting watermarking information by voting selection is proposed.The algorithm determines the embedded watermarking information according to the principle of minority obeying the majority,which effectively improves the robustness of the watermarking system.A genetic algorithm(GA)is proposed for voting to select watermarking information in DCT domain.The robustness and transparency of the optimized algorithm are improved compared with those of the original one.In the Discrete Wavelet Transform(DWT)domain,a robust watermarking algorithm combining DWT with Singular Value Decomposition(SVD)in HSI color space is proposed.GA is used to select the optimal embedding strength in H,S and I planes and set the threshold of peak signal-to-noise ratio(PSNR),which makes the algorithm robust under the condition that PSNR is greater than 35.A blind watermarking algorithm in wavelet domain based on SVD matrix analysis is proposed.By analyzing the characteristics of the unitary matrix after SVD decomposition,the algorithm modifies some coefficients of the unitary matrix according to the value of the watermarking sequence to extract the watermarking effectively.Finally,the algorithm is simulated and compared.Transparency is measured by the value of PSNR and robustness is measured by the value of Normalized Correlation(NC).The experimental results show that the proposed watermarking algorithm is feasible and robust to cropping,filtering,scaling,rotation and noise attacks.