The invention discloses an
atrial fibrillation monitoring method based on ECG signals, which relates to the technical field of
electrocardiogram analysis, the
atrial fibrillation monitoring method comprises the following steps: describing electrocardiogram 12 lead ECG original signals on the same canvas, and storing as an
ECG signal picture; marking a waveform and an R wave peak in the ECG signalpicture, and training a preset identification model by using a deep
convolutional neural network to obtain an R wave identification model; finding an R wave peak according to the R wave position, obtaining a position sequence vector V according to the R wave peak position, and then obtaining a
difference vector V1 of the R-R interval; performing marking according to the
difference vector V1, if
atrial fibrillation is judged, making the atrial
fibrillation as 1, otherwise, making the atrial
fibrillation as 0, and then obtaining a judgment
result set; performing classification training on the
result set by using a classifier to obtain an atrial
fibrillation recognition model; using the R-wave recognition model and the atrial fibrillation recognition model for recognizing the new electrocardiogram 12-lead ECG original signals, and then monitoring whether atrial fibrillation exists in the electrocardiogram or not. The method has the
advantage of being capable of rapidly judging whether atrial fibrillation exists or not.