中文摘要：

提出了基于云模型的音频水印技术。本丈利用云模型的随机性和模糊性，设定云模型的参数作为密钥，通过正向云发生器产生水印信息，水印映射到坐标轴上是一幅云图。采用小波域自适应量化的方法，根据嵌入位置的不同选择不同的量化系数，将水印嵌入到载体音频中，提取时既可以对云进行视觉上的判断，又能通过逆向云发生器得到云模型的参数，并且给出了定量判断的方法。实验结果表明，基于云模型的音频水印具有较高的鲁棒性和不可感知性，可有效保护音频作品的版权。

英文摘要：

A novel audio watermarking based on cloud model for copyright protection is presented. From the perspective of novel watermarking frame, the new watermarking is generated according to the given parameters of the cloud model as secret key, which is different from traditional watermarking such as binary images and pseudorandom sequences. The watermarking is embedded in the audio by the method of adaptive quantization, and the quantization coefficients are chosen diversely according to the different embedded locations in wavelet domain. After extracting the watermarking, it can be not only identified directly, but also can be gotten the parameters of the cloud model by reverse clouds generator The method of ration judgment is also given in the paper. Experiment results show that audio watermarking based on cloud model has better imperceptibility and robustness than the traditional watermarking, and can effectively protect the copyright of music works.