A Zero Watermarking Algorithm for Remote Sensing Images Based on NMF and Hu Moments
Aiming at the demand of copyright protection for remote sensing image data with strong cropping resistance,this paper proposes a zero watermarking algorithm based on the nonnegative matrix factorization(NMF)and Hu moments.The algorithm first decomposes the original image into coefficient matrix and basis matrix,then calculates the overall Hu moment invariant mean value and chunk invariant mean value,and binarizes these feature information and performs the dissimilarity operation with the copyright information to generate the zero watermark.The coefficient matrix as well as the zero watermark need to be deposited in a third-party organization.When detecting copyright information,the original data is back calculated from the data to be detected by utilizing the recoverable characteristics of NMF,and the zero watermark is extracted for copyright information comparison.The experimental results show that the proposed algorithm has uniqueness in copyright protection and robustness,especially in the resistance to strong cropping,showing its application value in the field of copyright protection of remote sensing image data.
万方数据
维普
