The Research Of Digital Image Watermarking Technology Based On Compressed Sensing

With the popularization of the computer,the use of a large number of the internet and the continuous development of multimedia,people can be rapidly transferred the digital information to all over the world through various kinds of electronic equipment;In the enjoyment of the quick service at the same time,people also put the safety of transmission more and more attention on the digital information.In the situation,an new information security technology--digital watermarking comes,and has the rapid development in the copyright protection of the multimedia information and the integrity authentication.But so far,the attack resistance and anti extraction capability of the digital watermarking is still limited.The randomness of the observation matrix and irreversible reconstruction algorithm in the compressed sensing has opened a new direction for the study of digital watermarking.The paper focuses on the research of digital watermarking technique based on compressed sensing theory,as follows:Based on the principle of digital watermarking technique,combined with the theory of compressed sensing,first of all,several classical reconstruction algorithms--OMP,ROMP,SAMP algorithm are analyzed by the two-dimensional image simulation.On this basis,this paper improves an RAMP algorithm.The algorithm controls the reconstruction accuracy by variable steps,is carefully detected in the small range of sparse signal values,is roughly detected in the large range of sparse signal values,and solves the under estimation and over estimation problems of caused by the unknown sparsity and the fixed step.And through the comparative analysis of image reconstruction effects,verifies the sampling rate is 0.7,the performance of RAMP algorithm is higher than these OMP algorithm,ROMP algorithm and SAMP algorithm in the terms of PSNR and T.Based on the RAMP algorithm,this paper puts forward a digital watermarking algorithm based on compressed sensing.The algorithm uses the Gauss random matrix to observe high-frequency coefficients of the watermarking,then high-frequency coefficients obtained by RAMP algorithm and original low frequency coefficients are embedded into DWT transform coefficients of the carrier in the multi resolution and the human visual system,at last,uses RAMP algorithm to reconstruct the watermarking.The experimental results show that: compared with OMP watermarking preprocessing algorithm,PSNR increases 2dB-4dB on average,NC value increases about 0.1 on average,has betterrobustness and imperceptibility.Finally,this paper achieves a digital watermarking system based on compressed sensing RAMP algorithm.The system realizes watermarking embedding,watermarking extraction,performance testing and other functions in the VC++6.0 platform;verifies that the digital watermarking system based on the improved RAMP algorithm can be effectively embedding and extracting watermarking;can adjust the embedding strength of the watermarking,resolves problems between invisibility and robustness of the digital watermarking;has the high security.