I recognise my code, however that version was not intended to work with a signal such as yours.
Try this instead —
load sig_ecg.mat
Fs = 300;
L = numel(sig);
t = linspace(0, L-1, L)/Fs;
% t=linspace(0,length(sig)/fs,length(sig));
figure()
plot(t,sig/max(sig));
grid
title('Original Signal in Time Domain');
xlabel('Time(s)');
ylabel('Amplitude');
Fn = Fs/2; % Nyquist Frequency (Hz)
NFFT = 2^nextpow2(L)
FTsig = fft((sig-mean(sig)).*hann(L), NFFT);
Fv = linspace(0, 1, NFFT/2+1)*Fn;
Iv = 1:numel(Fv);
figure
plot(Fv, abs(FTsig(Iv))*2)
grid
xlabel('Frequency (Hz)')
ylabel('Magnitude')
% xlim([0 5])
sig = bandstop(sig, [48 52], Fs, 'ImpulseResponse','iir'); % Remove 50 Hz Mains Interference
% figure()
% sig=sig-mean(sig); % Per visualizzare meglio lo spettro
% f=linspace(-fs/2,fs/2,length(sig));
% plot(f,fftshift(abs(fft(sig))));
% grid
% title('Signal in Frequency Domain');
% xlabel('Frequency (Hz)');
% ylabel('Amplitude');
% % Lo spettro del segnale mostra un consistente picco a 42Hz che è probabilmente correlato ad una "power line interference", dovuta ad artefatti da strumentazione. L'approccio più semplice per rimuovere quest'interferenza è di applicare un filtraggio IIR stop-band centrato a 42Hz. Inoltre notiamo la presenza di una baseline wander.
% F = sig;
% f = gca;
% EKG = findobj(f, 'Type','line');
% t = EKG.XData; % Recover Data
% s = EKG.YData;
% L = numel(t); % Signal Length
% Ts = mean(diff(t));
% St = std(diff(t)); % Test For Uniform Sampling
% tr = linspace(min(t), max(t), L); % New, Regularly-Sampled Time Vector
% Tsr = mean(diff(tr)) * 1E-3; % Convert To Seconds
% sr = resample(s, tr); % Resample To Regularly-Sampled Signal
% Fsr = 1/Tsr; % Sampling Frequency (Hz)
% Fnr = Fsr/2;
% figure
% plot(tr, sr)
% grid
% FTsr = fft(sr-mean(sr))/L; % Fourier Transform Of Mean-Corrected Signal
% Fv = linspace(0, 1, fix(L/2)+1)*Fnr;
% Iv = 1:numel(Fv);
% figure
% plot(Fv, abs(FTsr(Iv))*2)
% grid
% Fs = Fsr; % Sampling Frequency (Hz)
Fn = Fs/2; % Nyquist Frequency (Hz)
Ws = 0.5/Fn; % Passband Frequency Vector (Normalised)
Wp = 2.5/Fn; % Stopband Frequency Vector (Normalised)
Rp = 1; % Passband Ripple (dB)
Rs = 150; % Stopband Attenuation (dB)
[n,Wp] = ellipord(Wp,Ws,Rp,Rs); % Calculate Filter Order
[z,p,k] = ellip(n,Rp,Rs,Wp,'high'); % Default Here Is A Lowpass Filter
[sos,g] = zp2sos(z,p,k); % Use Second-Order-Section Implementation For Stability
sig_filtered = filtfilt(sos,g,sig); % Filter Signal (Here: isrg)
figure
freqz(sos, 2^14, Fs) % Bode Plot Of Filter
set(subplot(2,1,1), 'XLim',[0 50]) % Optional, Change Limits As Necessary
set(subplot(2,1,2), 'XLim',[0 50]) % Optional, Change Limits As Necessary
% figure
% plot(tr, sr_filtered)
% grid
% title('Signal in Frequency Domain with removal of baseline wander');
% xlabel('Frequency (Hz)');
% ylabel('Amplitude');
figure
plot(t, sig_filtered)
grid
title('Filtered Signal in Time Domain');
xlabel('Time(s)');
ylabel('Amplitude');
This removes the baseline drift, and also eliminates the 50 Hz mains interference with a separate filter. The bandstop function was introduced in R2018a, so you should have it. You can also substitute my filter code with a highpass call, similar to the syntax I used in the bandstop call (although only the 0.5 Hz cutoff frequency, not the 2.5 Hz passband frequency, would need to be provided).
I commented-out parts of the code that are either not appropriate or not helpful for your signal, since this was originally intended to recover data from a .fig file with a time vector that did not have constant, uniform sampling intervals. (I am not certain why I used a two-sided Fourier transform, since I usually use one-sided Fourier transforms for signal processing problems, as I have in revising this.)
.
