图像通常包含许多边缘或噪声点导致二进制小波域中高频成分较多,而传统二元树复小波算法存在不适于处理高频成分缺点.为此,利用多频带(M波段)正交小波能够更好的将能量压缩为窄带信号高频分量特点,对传统二元树复小波算法结构进行改进,并结合分数傅立叶算法和SVD算法实现数字水印高频信号的高精度处理,设计了分数傅立叶M波段二元树复小波SVD数字水印算法.仿真结果显示,改进算法提高了算法对高频压缩攻击、滤波攻击的鲁棒性.“,”The image usually contains a lot of edges, which lead the binary wavelet domain have many high-frequency components, meanwhile, the traditional dual-tree complex wavelet algorithm is not suitable for processing high frequency components. For this reason, here makes use the multi-band (M-band) orthogonal wavelet with its advantage in compressing energy to narrowband signal frequency component to improve the traditional dual-tree complex wavelet algorithm, and combined with the fractional Fourier transform and SVD algorithm to realize the high-precision processing for high-frequency signals of digital watermark, then, the fractional FourierM-band dual-tree complex wavelet and SVD based digital watermarking algorithm is proposed. Simulation results show that the algorithm can improve the robustness of the algorithm against high frequency compression attacks and filtering attacks.