| Abstract: | The mixed noise in the acquisition process of pulse wave signals will affect the signal analysis, how to effectively eliminate the noise and complete the pulse wave analysis has important practical significance. In this paper, empirical mode decomposition(EMD) and ensemble empirical mode decomposition(EEMD) were used to realize scale decomposition of pulse wave signals to obtain intrinsic mode function(IMF). A band-pass filter was implemented according to the characteristic time scale parameters of the IMF. After filtering and reconstruction, the pulse wave denoising was completed. The denoising effects of EMD, EEMD and wavelet transform were compared in terms of mean square error and signal-to-noise ratio. The result shows that EMD and EEMD are better than wavelet transform, and the effects are similar. Further comparing the Hilbert-Huang spectrum of EMD and EEMD, it can be seen that EEMD can not only avoid mode mixing, but also facilitate the analysis of pulse wave signals. |
