EEMD-PCA-BASED METHOD AND DEVICE FOR REMOVING MOTION ARTIFACT FROM EEG SIGNAL

The present invention relates to the field of EEG mobile electroencephalogram processing. Disclosed are an EEMD-PCA-based method and device for removing a motion artifact from an EEG signal, comprising: decomposing a single-channel EEG signal on the basis of EEMD to obtain an intrinsic mode function of each order (S101); separating the intrinsic mode function of each order on the basis of PCA to obtain principal components (S102); calculating a self-correlation of each principal component (S103); determining, as an artifact principal component, the principal component of which the self-correlation is greater than a preset threshold (S104); removing the principal component which is determined as an artifact (S105); and performing PCA inverse change processing on the remaining principal components and then performing EEMD inverse change processing on same, to obtain an EEG signal subjected to noise removal (S106). Compared with the prior art, the artifact removal method can significantly improve the artifact re