基于Arnold变换和DWT彩色图像数字盲水印算法

Blind Digital Watermarking Algorithm for Color Image Based on Arnold and DWT

针对目前大多数彩色图像数字水印算法中嵌入的水印仍是二值图像或灰度图像,并且水印提取过程只实现了水印图像的提取而不能恢复原始宿主图像的问题,提出了一种将彩色水印嵌入到彩色宿主图像中的盲水印算法。水印提取过程无需原始宿主图像和原始水印图像,仅利用嵌有水印的图像就能同时恢复原始水印图像和原始宿主图像。该算法对彩色水印的R、G、B分量分别进行不同次数的Arnold变换,以达到置乱的目的;然后对彩色宿主图像的R、G、B分量分别进行离散小波变换(DWT:Discrete Wavelet Transform),取出各自的水平细节系数,最后把置乱后的水印R、G、B分量对应地嵌入到宿主图像的R、G、B水平细节系数中,嵌入方式采用二维离散小波逆变换(IDWT:Inverter Discrete Wavelet Transform)。实验表明,该数字水印算法具有很好的透明性和安全性,同时对诸如涂抹、叠加噪声、剪切、JPEG压缩等常见攻击具有很好的鲁棒性。

Considering most of the present color image watermarking techniques still use binary images or gray images as the watermarks,and these techniques only extract the watermarks while ignoring the original host image restoration.A blind digital watermarking algorithm is proposed to embed a color watermark image into a color host image.The algorithm allows extracting watermark without the original watermark and host image.The original watermark image and host image can be restored simultaneously by using the image embedded with watermark.Firstly,Arnold transform was respectively used on R,G,B component of the watermark image with different times to scramble the watermark image.Then decomposed separately the R,G,B component of the color host image by 2D DWT（Discrete Wavelet Transform）,and got the respective horizontal detail coefficients.Finally embedded the disorderly R,G,B component of the watermark image into the horizontal detail coefficient of R,G,B component of the host image.IDWT（Inverter Discrete Wavelet Transform） is masterly used as the embed way.Due to using HVS（Human Visual System） and adopting multiple key mechanisms,the digital watermarking algorithm has very good transparency and security.In addition,the algorithm is robust against various signals processing such as scrawling,adding noise,cutting,and JPEG compression etc.