Research of Intrinsic Mode Function Illusive Component Recognition and Adaptive Threshold Denoising Method Based on Empirical Mode Decomposition

Empirical mode decomposition (EMD) has been widely used since it was proposed. However, there are illusive component in intrinsic mode function (IMF) obtained by EMD. In addition, the demarcation point K of signal and noise modal function in traditional EMD denoising method are difficult to select accurately. In order to solve the above problems, the normalized energy distribution characteristics of IMF components are used to identify and eliminate illusive components firstly. Then, an adaptive threshold optimization method to denoise IMF components is presented in this paper. Finally, the IMF components after noise reduction are superimposed and merged to obtain the true value of the signal. Through the analysis of examples, it is found that comparing with wavelet transform and traditional EMD denoising methods, this method has a great improvement in signal-to-noise ratio and noise suppression ratio.