Based On Empirical Mode Decomposition The Wavelet Threshold Signal Denoising Research,

Signal is the carrier of the information we get, but in practice a variety of signals collected, for the reasons of the data acquisition environment and collecting instruments, inevitably doping some noise component which we want nothing to do with the signal, so in many cases. the signal denoising becomes very necessary and very meaningful.Denoising methods are many, but each method has its own advantages, disadvantages and application conditions. Previously, one of the most commonly used is filtering denoising method based on Fourier transform, but this method does not apply to non-stationary signals. Currently, the most studied signal denoising methods, the most commonly used are Denoising Based on Wavelet Transform and Denoising Method Based on EMD. There are also the joint denoising method of EMD and wavelet transform, this study also is the wavelet threshold denoising method based on EMD, first introduced the wavelet transform theory and the theory of EMD, the paper focused on several problems which encountered in the wavelet threshold denoising method, one is how to choose the appropriate wavelet and wavelet decomposition level, the second is how to select the threshold and threshold function adopted. The two issues must be addressed in the wavelet thresholding denoising and the settlement directly affects the denoising effect. Finally, using the denoising methods studied in this paper compare with the existing denoising methods, obtained the superiority of the denoising method in text.The idea of denoising method proposed in this paper is. first of all do EMD for the noise signal, by the order IMF component, and then according to the conclusion that the multiplied of energy density and its corresponding average cycle is a constant of the order IMF identify in previous IMF component is the noise component.direct interception off,the signal is the main ingredient of remaining IMF component, but still contains noise.and then the remaining components of the IMF do wavelet thresholding denoising.and finally with the IMF components of the denoised signal can be reconstructed.