：由于许多传统的去噪方法在强背景噪声情况下提取声音信号的能力变弱甚至失效， 提出

应用独立成分分析( I C A) 方法对声音信号进行特征提取， 并证明了这种 I C A 变换能增强语音和音

乐信号的超高斯性． 在此基础上， 应用 I C A基函数作为滤波器， 通过阈值化的去噪方法对含有强高

斯背景噪声的声音信号进行去噪仿真实验． 结果表明， 本方法明显优于传统的均值滤波和小波去噪

方法， 为强背景噪声下弱信号的检测提供 了新的途径．-: As many of the traditional de-noising method in case of strong background noise, the ability to extract the voice signal even weaker failure, the application of independent component analysis (ICA) method of voice signal feature extraction, and prove that this transformation can be enhanced voice ICA and music of super-Gaussian signals. On this basis, the application of ICA basis function as a filter, through the threshold of the denoising method of Gaussian background noise contains strong voice signal denoising simulation. The results show that this method is obviously superior to the traditional mean filtering and wavelet denoising methods for the strong background noise under the weak signal detection provides a new way.