改进压缩感知的舰船数字图像水印算法

Watermarks for ship digital images based on improved compression sensing

为提升数字图像水印嵌入与提取效果,研究改进压缩感知的舰船数字图像水印算法。通过离散小波变换稀疏分解舰船数字图像,获取高频与低频系数矩阵。利用Kent混沌观测矩阵改进压缩感知理论,依据改进压缩感知理论获取高低频系数矩阵观测值。在高频系数矩阵观测值内嵌入水印,并记录嵌入位置,建立索引矩阵。通过压缩感知重构的改进正交匹配追踪算法,重构嵌入水印后的高低频系数矩阵观测值,获取高低频系数矩阵。融合高低频系数矩阵,并离散小波反变换处理融合后的矩阵,获取嵌入水印的舰船数字图像。利用索引矩阵与高频系数矩阵的观测矩阵,提取水印。实验证明:该算法可有效嵌入与提取舰船数字图像水印;在不同攻击类型时,该算法水印嵌入的透明性均较优,水印提取的精度均较高,具备较优的抗旋转与抗剪切攻击能力。

In order to improve the embedding and extraction effect of digital image watermarking,an improved ship digital image watermarking algorithm based on compressed sensing is studied.The high frequency and low frequency coefficient matrices are obtained by sparse decomposition of ship digital images by discrete wavelet transform.Using Kent chaotic observation matrix to improve compressed sensing theory.The observation values of high and low frequency coefficient matrices are obtained according to the improved compressed sensing theory.The watermark is embedded in the observed value of the high frequency coefficient matrix,and the embedding position is recorded to establish the index matrix.Through the improved orthogonal matching pursuit algorithm of compressed sensing reconstruction,the observed values of high and low frequency coefficient matrices embedded in the watermark were reconstructed,and the high and low frequency coefficient matrices were obtained.The high and low frequency coefficient matrices are fused,and the fused matrix is processed by inverse discrete wavelet transform,and the watermark embedded ship digital image is obtained.Using index matrix and observation matrix of high frequency coefficient matrix,the watermark is extracted.Experimental results show that the algorithm can effectively embed and extract ship digital image watermarks.Under different attack types,the transparency of watermark embedding is better,the precision of watermark extraction is higher,and the algorithm has better anti-rotation and anti-shear attack ability.