摘要
Considering the properties of slow change and quasi-periodicity of magnetocardiography （MCG） signal, we use an integrated technique of adaptive and low-pass filtering in dealing with two-channel MCG data measured by high Tc SQUIDs. The adaptive filter in the time domain is based on a noise feedback normalized least-mean-square （NLMS） algorithm, and the low-pass filter with a cutoff at 100Hz in the frequeucy domain characterized by Caussian functions is combined with a notch at the power line frequency. In this way, both relevant and irrelevant noises in original MCG data are largely eliminated. The method may also be useful for other slowly varying quasi-periodical signals.
Considering the properties of slow change and quasi-periodicity of magnetocardiography （MCG） signal, we use an integrated technique of adaptive and low-pass filtering in dealing with two-channel MCG data measured by high Tc SQUIDs. The adaptive filter in the time domain is based on a noise feedback normalized least-mean-square （NLMS） algorithm, and the low-pass filter with a cutoff at 100Hz in the frequeucy domain characterized by Caussian functions is combined with a notch at the power line frequency. In this way, both relevant and irrelevant noises in original MCG data are largely eliminated. The method may also be useful for other slowly varying quasi-periodical signals.
