Algorithm Research On EGG Denoising And R Wave Detection

Electrocardiogram(Electrocardiograph,ECG)is a graphic technology that records the change of electrical signals produced by the beating cycle of the heart by electrocardiograph.It can be used to record the electrical activity of the normal heart and diagnose various heart diseases.However,the electrical signals at the center of the acquisition process are often subject to various disturbances.Therefore,the denoising and processing of the ECG signals and the detection of the characteristic waveforms have always been the research focus in the field of medical signal processing.This dissertation describes the research status of ECG signal denoising and R wave detection,and improves the existing ECG signal denoising algorithm and R wave detection algorithm,and achieves satisfactory results through simulation experiments.The main work of this dissertation is as follows:(1)Firstly,the generation mechanism,basic components and the physiological characteristics of the various characteristic waveforms,the parameters contained in the wavebands,and the types of interference noise received by ECG signals are introduced.ECG signal is a kind of strong randomness,weak amplitude and low signal-to-noise ratio of physiological signals.During the acquisition,ECG signals are easily affected by the factors such as electrocardiogram,electrode connection and the motion artifact of the patient themselves.These common noises mainly include EMG noise,frequency noise and baseline drift.(2)According to the different properties of all kinds of noise,different noise removal methods are adopted for different noise.In view of the removal of baseline drift,a combination of mathematical morphology and Hilbert vibration is presented in this dissertation.First,the QRS wave group is separated by improved morphological open and close operation,and the P wave and T wave with baseline drift are retained.Then the baseline drift is extracted by Hilbert’s vibration.Finally,the QRS wave packet and the P and T waves filtered by the base drift are superimposed to get the undistorted ECG waveform.For the frequency noise and EMG noise,an algorithm based on the ensemble combination of empirical mode decomposition and new wavelet threshold is presented in this dissertation.First,several intrinsic mode function(IMFs)of ECG signals are obtained by empirical mode decomposition,and the number of noisy IMFs is determined,and then the noisy IMFs component are decomposed by wavelet,the noise of low scale(such as the first scale)is removed by adaptive double threshold,and the noise of the lower scale(such as the third scale)is removed by new threshold function.Finally,the de-noised IMF component and the signal IMF component are reconstructed to obtain a clean ECG signal.In this dissertation,we use Matlab2015a software programming to verify and simulate the data in MIT-BIH arrhythmia database,and the result is good.(3)Shannon energy(Shannon energy)is applied to the R wave detection algorithm.The algorithm uses the band-pass filter to filter the ECG signal,then the ECG signal denoising of the first-order difference and amplitude normalized;then the Shannon energy calculation,window function smoothing for signal processing,to extract the envelope,prominent R peak;then through a series of nonlinear operation and corrected,finally detect the R wave peak position.In this dissertation,the Matlab2015a software programming is used to simulate the data of the MIT-BIH arrhythmia database.It is proved that the algorithm can not only overcome the shortcomings of threshold selection,poor real-time performance and large computation,but also has good detection precision.