2. Physiological signal processing - 1. ECG signal (including Matlab code and data)

This article is suitable for quickly understanding ECG signals and can perform data filtering.

1. ECG data preprocessing (eliminating power frequency interference and baseline drift)

* Please see the bottom of the article for ECG data and rdmat function

1. Import ECG data

## 心电图导入及读取
clc;
[TIME,M,Fs,siginfo]=rdmat('100m');# 通过读取函数ramat对心电图进行处理
Fs=1500;# 采样频率
plot(TIME,M);

2. Select the first 1000 data points for subsequent processing

ecg = M(1:1000);
TIME = TIME(1:1000);
plot(TIME,ecg)

3. Remove the 50Hz frequency interference in the ECG signal (usually 50Hz in China) - notch filter

## Butterworth 陷波滤波器-去除工频干扰
# 设计滤波器
d = designfilt('bandstopiir','FilterOrder',2, ...
               'HalfPowerFrequency1',49,'HalfPowerFrequency2',51, ...
               'DesignMethod','butter','SampleRate',Fs);
#应用滤波器去除50Hz干扰
ecg_50 = filtfilt(d,ecg);

4. Remove low-frequency signals (below 5Hz) from the ECG signal - eliminate baseline drift

##bandpass滤波器-解决基线漂移
fmaxd_1=5;# 截止频率为5Hz
fmaxn_1=fmaxd_1/(Fs/2);
[B,A]=butter(1,fmaxn_1,'low');
ecg_low=filtfilt(B,A,ecg_50);# 通过5Hz低通滤波器的信号
ecg1=ecg_50-ecg_low; # 减去5Hz低频信号
plot(TIME,ecg1)
xlabel('t(s)');
ylabel('mv');
title('期望信号');

5. Add random noise

* wgn function: randomly add white noise

##添加随机噪声信号
Noise_White = (0.1*wgn(1,length(TIME),2))'; 
plot(Noise_White)
title('噪声信号');

6. Signal after noise interference

##噪声干扰后信号
Mix_Signal = ecg1 + Noise_White;
plot(TIME,Mix_Signal)
xlabel('t(s)');
ylabel('mv');
title('噪声干扰后信号');

2. Filtering (Wiener filtering)

*h is the value obtained after running Wiener filter, which can be obtained by running the Wiener filter design part below :

1. Signal processing - 3. Wiener filtering (including matlab code)_Quite interesting_'s blog-CSDN blog

##维纳滤波
Signal_Filter = filter(h,1,Mix_Signal);# 将输入信号通过维纳滤波器
figure(1)
plot(TIME,Signal_Filter)
xlabel('t(s)');
ylabel('mv');
title('维纳滤波后信号');

3. Mean square error

* The statistical mean square error of the signal after introducing noise relative to the original signal:

mse1 = 0.0153

* Statistical mean square error of the filtered signal relative to the original signal:

mse2 = 0.0069

After applying Wiener filtering, the signal is closer to the original signal.

* ECG data and rdmat function

Link: https://download.csdn.net/download/qq_41958946/86841684

Original source of ECG data: MIT-BIH Arrhythmia Database (mitdb)