Research On Color Image Digital Watermarking Algorithm Based On QR Code

With the advancement of digitization and networking,the security of digital media and the problem of copyright protection have become increasingly prominent.Digital watermarking as an effective solution for copyright protection has become a research hotspot.But so far there is no universal digital watermarking algorithm that is resistant to comprehensive attacks.Based on the research of the existing digital watermarking technology,this paper combines the excellent characteristics of DCT transform,singular value decomposition,NSST and Hessenberg decomposition to study the more robust color image QR code digital watermarking algorithm.The main work of the paper is as follows:1.In order to improve the robustness of the watermarking algorithm against all-round attacks,a multi model watermarking algorithm based on QR code is designed.The algorithm performs two DCT transformations on the L component of the CIELab space of the carrier image,and then selects the first three low-frequency coefficients at the same position of each sub-block to construct multiple matrix models to be embedded,and multiple encrypted QR Code watermark copies are embedded in the largest diagonal elements after SVD decomposition of each model.In the algorithm,the watermark adopts the dual encryption method of Arnold and bitmap segmentation XOR to enhance the security of the watermark.The optimal embedding strength is selected according to the experimental test method,and finally the watermark is extracted by multi-model statistical voting,which enhances the algorithm stability.2.Aiming at the problem of invisibility and robust balance in traditional embedded watermarking algorithm,a zero watermarking algorithm based on NSST and Hessenberg decomposition is proposed.The algorithm first performs NSST transformation on the L component of the image CIELab space,then performs DCT and block processing on its low-frequency components,and uses the maximum value of Hessenberg decomposition of each sub-block to construct a feature matrix,and finally XORs the feature matrix and copyright information Zero watermark,copyright information is obtained through XOR of the saved zero watermark and feature matrix during copyright authentication.The matrix decomposition calculation used by this algorithm has lower computational complexity and does not modify the original image.3.Experiments and tests were carried out on the two proposed digital watermarking algorithms,and compared with similar algorithms.The experimental results prove that the two watermarking algorithms in this paper have good resistance to common noise,filtering,scaling,rotation,JPEG compression,and cutting attacks.Compared with similar algorithms,they are more effective in noise,cutting,and rotation.And the advantage of robustness in compression attacks is significant.