基于DWT的彩色图像绿色分量数字水印方案

DWT-based color image watermarking scheme of embedding watermark into the green components

提出了一种在彩色图像的绿色分量嵌入置乱水印的方案,嵌入过程在小波变换域进行。原始二值图像水印经过置乱加密,并用均值为零、相互正交的两个伪随机序列扩频后,嵌入到绿色分量经两层DWT变换后的低频子带系数中。选取低频子带中能量最大的系数嵌入水印,并根据低频系数的能量比来确定水印嵌入的强度。针对水印嵌入方案,设计了有源提取水印的算法和水印盲检测算法。实验结果证明,该水印方案能有效地抵抗加性噪声(如椒盐噪声和高斯噪声)、低通滤波、缩放、图像剪切、行(或列)删除、JPEG压缩等各种攻击,具有很强的顽健性。

The scheme of embedding a scrambled watermark into the green component of the color image was proposed.The embedding process was implemented in the discrete wavelet transformation（DWT） domain. The original binary watermark image was firstly encrypted through scrambling technique, and then spread with two orthogonal pseudo-random sequences whose means were equal to zero, finally embedded into the DWT low frequency sub-band of green components. The coefficients whose energies were larger than the others were selected to hide watermark, and the hided watermark strength was determined by the energy ratio between the selected coefficients energies and the subband＇s mean energy. The retrieving watermark algorithm and the blind detecting algorithm were designed in accordance with the embedding scheme. The experimental results demonstrate that the proposed watermarking scheme is very robust to the attacks such as additive noise, low-pass filtering, scaling, cropping image, rows（or column） deleting and JPEG compression.