Abstract: Firstly, using fractional Fourier transform algorithm, the color image was transformed in fractional frequency domain, the frequency domain information of the image was analyzed, and the coefficients suitable for embedding watermark were selected as the embedding position of watermark. Secondly, the concealment characteristics of the frequency domain coefficients were defined, and the concealment characteristics were classified, so as to achieve the optimal watermark embedding coefficient. Finally, the lifting wavelet transform and compressive sensing measurement of binary watermark images were made, and the measurement results were embedded into the carrier images as watermark information. It solved the problem of high redundancy of embedded information in most current watermark algorithm and insufficient antiattack ability of the watermark image. Experimental results show that the proposed algorithm has good performance in resisting geometric attack and noise interference, it significantly reduces the redundancy of watermark information, improves the efficiency of embedding and security of watermark information, and increases the reliability of watermark algorithm.

Key words:  , watermark adaptive embedding, fractional Fourier transform, lifting wavelet transform, compressive sensing measurement