论文部分内容阅读
经验模式分解(EMD)是一种自适应的非线性、非平稳信号分析方法,广泛用于非参数化信号去噪。由于存在边界效应和模式混叠现象,EMD去噪效果受到一定影响。为了提高去噪性能,文章提出一种基于区间软阈值和部分重构的多维经验模式分解(MEMD)去噪方法。该方法利用MEMD对于高斯白噪声的良好二进滤波特性,以固有模式函数(IMF)与输入信号概率密度函数(PDF)之间的相似度来选择最佳模式函数。根据IMF本身特点,采用区间软阈值去噪方法对选取的IMF分量进行去噪,最后结合部分重构实现信号的去噪。实验仿真和实测EEG信号处理结果表明,与小波变换和EMD-CIIT方法相比,文中方法对单通道信号中高斯噪声具有2dB~3dB的性能提升,同时还可以对多通道信号进行联合去噪,是一种有效的信号去噪新方法。
Empirical Mode Decomposition (EMD) is an adaptive, non-linear, non-stationary signal analysis method that is widely used for denoising non-parametric signals. Because of the boundary effect and mode aliasing, the effect of EMD denoising is affected. In order to improve the performance of denoising, a novel multi-dimensional empirical mode decomposition (MEMD) denoising method based on interval soft threshold and partial reconstruction is proposed. This method uses the good binary filtering characteristic of MEMD for Gaussian white noise and selects the best mode function based on the similarity between the intrinsic mode function (IMF) and the input signal probability density function (PDF). According to the characteristics of the IMF itself, the interval-based soft threshold denoising method is used to denoise the selected IMF components, and the signal denoising is finally combined with the partial reconstruction. Experimental results show that compared with wavelet transform and EMD-CIIT method, the proposed method has the performance improvement of 2dB ~ 3dB for Gaussian noise in single-channel signal, meanwhile it can also combine the multi-channel signal to denoise, Is an effective new method of signal denoising.