Study Of FECG Signal Processing Based On Wavelet Transform And Non-negative Matrix Factorization

FECG signal is the primary signal generate by the heart during fetal development,reflecting the entire process and changes of heart activity.Through the analysis of FECG,can know the development of the fetus intrauterine condition in time,judge whether there is abnormal defects.However,the signals collected from the abdominal wall of the mother include ECG,frequency interference,baseline drift,and various other noises.The FECG itself is a low-frequency and weak non-stationary signal.How to extract a clearer FECG signal from the abdomen signal has been a research hotspot in the field of biomedical signal processing.This article describes the FECG related knowledge and commonly used processing methods,in the premise of summing up the excellent experience,through the experimental comparison and analysis,put forward a method for FECG signal processing based on wavelet transform and the decomposition of nonnegative matrix factorization.At first,Preprocessing the mixed signal,Using the comb filter is designed to eliminate 50 Hz power frequency interference and its harmonics,the low-pass filter removes high frequency noise,and the median filter removes the baseline drift.And verify the pretreatment effect.Then,Introduce a maternal cardiac electrical signal separation method based on wavelet transform.Wavelet transform is widely used in signal processing because of its multiresolution characteristics and local feature representation ability.According to the characteristics of the ECG signal to choose the appropriate wavelet and wavelet transform decomposition layers.To get the approximate signal threshold detection,and get all R wave peaks of ECG signal,Then the corresponding high frequency component of the ECG signal is intercepted from the detail signal,and then the ECG signal reconstructed by wavelet reconstruction.Using the filtered signal minus the reconstructed ECG,the FECG containing noise can be obtained.After that,put forward a method of FECG signal processing based on non-negative matrix factorization.When dealing with FECG signal containing noise in a single-channel model,in order to satisfy the requirement of nonnegative matrix decomposition,Firstly,the time-frequency transformation is performed to obtain the Born-Jordan distribution.By converting to the power spectrum in the frequency domain,an appropriate frequency band is selected for NMF decomposition to obtain FECG characteristic signal,Peak detection can be calculated to obtain the fetal instantaneous heart rate.Simulation tests and results analysis were performed by using simulated data and clinical data from the MIT-BIT ECG database to verify the reliability of the algorithm.The experimental results show that the algorithm proposed in this paper can effectively extract fetal ECG signals and calculate a more accurate fetal heart rate.It can be used as a reference for prenatal diagnosis of fetal abnormalities.