The invention aims at providing a
muscle disease monitoring method based on multi-channel surface
electromyography (sEMG). The method comprises the steps that preprocessing is conducted on multi-channel sEMG signals firstly, a first channel is selected as a reference, and differencing is conducted on the signals of other channels and the
signal of the first channel; then distribution moments are extracted through a K-means clustering
convolution kernel compensation (KMCKC) method, and a single waveform is extracted; finally, multiple features of the waveform are fused, and the
muscle state is evaluated. Due to the fact that external interference has influence on all electrodes, by meas of differencing between the
signal of the first channel and the sEMG signals of original multiple channels, the external interference is effectively reduced, subsequent detection results cannot be affected, and the accuracy of
muscle detection is improved. By means of application of the multi-feature parameters, the
instability of single-parameter monitoring is effectively avoided, and the monitoring robustness is promoted. By means of the
muscle disease monitoring method based on the multi-channel surface
electromyography (sEMG), the defects in the prior art are effectively overcome, and important application value is achieved.