基于小波包和改进EMD的脑电信号消噪研究

由于采集到的脑电信号含有噪声,提出了总体经验模态分解(EEMD)的希尔伯特黄变换(HHT)结合小波包分析的脑电信号去除噪声的方法。含噪脑电信号经EEMD分解可以得到一定数量的IMF分量,而且可以解决经验模态分解(EMD)时的模态混叠问题,然后,对IMF分量进行Hilbert变换,分析Hilbert谱,把含噪的IMF部分进行小波包处理,最后,把各IMF相加,可得处理过噪声的脑电。经验证得,单独使用小波包方法消噪和改进EMD消噪,都没有小波包结合改进EMD方法的信噪比高,提高了去噪效果,有利于更精确的诊断医学疾病。

Research on De-noising of EEG Based on Wavelet Packet and Improved EMD

Since the collected EEG signals contain noise,a method based on the Hilbert-Huang transform(HHT) of Ensemble Empirical Mode Decomposition(EEMD)combined with the wavelet packet is proposed to remove noise from EEG. The EEMD decomposition can get a certain number of IMF components of noisy EEG signals and solve the modal aliasing problem in empirical mode decomposition(EMD). Then,Hilbert transform is used to analyze IMF components,Hilbert spectrum is analyzed,IMF part of the wavelet packet processing,and finally,the IMFs are add-ed,have been dealt with noise EEG. It has been verified that using the wavelet packet method alone to denoise and improving the EMD denoising,no wavelet packet combined with the improved EMD method has a high sig-nal-to-noise ratio,which improves the denoising effect and is more accurate for the diagnosis of medical diseases.