detect QRS signal from ekg shield

my graduation project is to get ecg signal from arduino uno -ekg shield -3 electrode then detect QRS point to make some process on it , but there will be much noise in signal so is there a way to detect QRS points from this noise

Answer 1

[Christoph F.]

If the ECG is noisy, the first place to look for issues would be the circuitry and the connection to the patient. Electrode contact impedance, ground loops and similar issues that cause line noise, a driven right leg circuit for reducing common mode interference, etc. Even an unprocessed ECG signal should not be unrecognizably noisy.

Digital signal processing like filtering can remove some noise, but it can only do so much. It cannot remedy every problem that is out in the analog world.

[Christoph F.]

You could start with analyzing the noise content using the fft function or possibly the pwelch function from the signal processing toolbox if you have access to it.

Then, given some additional information like the sampling rate and the expected output bandwith, a filter can be designed that attenuates the noise.

[Star Strider]

A ‘healthy’ EKG can be processed with a bandpass filter of 1 Hz to 50 Hz, and in the presence of arrhythmias, 1 Hz to 100 Hz. (The 1 Hz lower cutoff is to eliminate motion artefacts and d-c offset.)

The sampling frequency should be at least 250 Hz.

[Christoph F.]

The standards usually specify a bandwidth of (at least) 0.67 Hz to 40 Hz for a monitoring ECG. Which would be a good starting point for R-wave detection, as an upper edge frequency of 40 Hz leaves enough room to attenuate line noise, which is usually 50 Hz or 60 Hz.

Diagnostic ECGs are more interesting, since the highpass filter is not specified by cutoff frequency, but by distortion of the S-T segment (a first-order 0.05 Hz digital filter fulfills the requirements, a second-order 0.05 Hz digital filter will not).