目的 为了进一步提高水印的鲁棒性和保护性，文中提出一种基于QR码和DWT-SVD分解的双重水印方案。方法 将次级水印的奇异值嵌入主水印经一级离散小波变换后低频系数的奇异值中，再将含有次级水印信息的主水印奇异值嵌入主图像的二级小波变换后低频系数的奇异值中。结果 经仿真实验发现，文中算法的主水印和次级水印PSNR值可达66 dB和60 dB，并对于高斯噪声、JPEG压缩、JPEG2000压缩、旋转攻击、对比度增强以及缩放攻击有一定的鲁棒性，攻击后提取的水印相关系数接近于1。结论 算法属于非盲水印，在多种攻击情况下都能检测到次要水印，但有时候主水印会严重失真。

The work aims to propose a dual watermarking algorithm based on QR code and DWT-SVD decomposition to further improve the robustness and safety of watermark. The singular value of the secondary watermark was embedded into the singular value of the low-frequency coefficient of the primary watermark after the first-order discrete wavelet transform and then the primary watermark singular value containing the secondary watermark information was embedded into the singular value of the low frequency coefficient after the second wavelet transform of the main image. Through the simulation experiment, the PSNR values of primary watermark and secondary watermark by the proposed algorithm were respectively 66 dB and 60 dB. The proposed algorithm was robust to Gaussian noise, JPEG compression, JPEG2000 compression, rotation attack, contrast enhancement and scaling attacks. The relevant watermark coefficient extracted after attack was close to 1. The algorithm is a non-blind watermark. The secondary watermarks can be detected under various attack conditions, but sometimes the primary watermark maybe seriously distorted.