Abstract: To enhance the watermark robustness and application，a new embedding strategy for the watermark is proposed based on DWT domain.The algorithm is proposed based on the influence of the selection for wavelet bases and the robustness of watermarking.In selects wavelet Haar base，according to watermark’s length and human visual characteristic，with wavelet transform，the image is decomposed into suitable classification，watermark components with different strength are inserted into different wavelet coefficients based on the results of classification.Experimental results show that the proposed algorithm is robust to some image processing operations and Joint Photographic Experts Group（JPEG）.

摘要： 为了使水印具有更强的鲁棒性和更广泛的适应性，提出了一种基于Haar小波基的离散小波变换（DWT）的隐形水印算法。该算法根据数字水印小波基的选择与水印鲁棒性的关系，选取Haar小波基，并根据水印长度和人眼视觉特性，自适应地把图像进行相应的小波级数分解，并对最低层逼近系数分类，根据分类的结果将不同强度的水印分量嵌入到低频系数中。实验结果表明，该算法具有不可见性，经过某些图像处理操作和JPEG有损压缩后仍是鲁棒的。