ECG denoising method based on EEMD-Hilbert transformation

Abstract

The invention discloses an EEMD (ensemble empirical mode decomposition)-Hilbert transformation-based ECG (ECG noise) denoising method, which comprises the following steps of: firstly, performing modal decomposition on original ECG noise by adopting an integrated empirical mode decomposition method to obtain a group of intrinsic mode functions consisting of a plurality of intrinsic mode functions, and then sequentially performing Hilbert transformation, phase solving and instantaneous frequency solving on the basis of the obtained group of intrinsic mode functions to obtain an ECG denoising result. Classifying each intrinsic mode function based on the central instantaneous frequency of a group of instantaneous frequencies corresponding to each intrinsic mode function, and abandoning the intrinsic mode function which is classified to be mainly based on baseline drift noise; respectively removing myoelectricity interference noise in the intrinsic mode function which is classified to be signal-dominated and the intrinsic mode function which is classified to be high-frequency noise-dominated, and reconstructing to obtain a final de-noised ECG; the method has the advantages that the baseline drift noise and the myoelectricity interference noise can be reduced more effectively, the obtained final de-noised ECG signal still keeps obvious ECG morphological characteristics, and the de-noising effect is good.

Description

technical field [0001] The present invention relates to an ECG denoising method, in particular to an ECG denoising method based on EEMD-Hilbert transform. Background technique [0002] During the acquisition of an ECG signal (Electrocardiogram signal, electrocardiogram signal), there will inevitably be baseline drift noise and electromyographic interference noise. However, the ECG signal collected by the wearable device will suffer from more serious noise interference, so that the ECG signal cannot be effectively used. [0003] In order to obtain the ECG signal with less interference and can be effectively utilized, the wavelet threshold denoising method is widely used in the field of ECG signal denoising. The wavelet threshold denoising method usually has the following five steps: step 1, use wavelet transform to decompose the noisy signal (ie the original ECG signal) into multi-level detail wavelet coefficients of different scales; step 2, respectively The multi-level de...

AI Technical Summary

Problems solved by technology

[0004] However, the above-mentioned wavelet threshold denoising method has limited effectiveness of the threshold evaluation method, and the spectral range of the EMG noise overlaps with that of the ECG signal. Denoising the overlapping part of the signal spectrum, the signal part may be regarded as noise, so that the denoised ECG signal will not only produce local distortion, but also the morphological characteristics of the ECG signal may be filtered out together as noise. This leads to poor final denoising effect

Images