Signal underdetermined blind source separation method and device based on time-frequency distribution

The invention discloses a signal underdetermined blind source separation method and device based on time-frequency distribution, belongs to the technical field of signal processing, and is used for solving the technical problem that the existing underdetermined blind source separation technology is poor in signal separation effect and cannot meet use requirements. The method comprises the following steps: constructing a linear hybrid model; wherein the linear mixing model comprises a linear instantaneous mixing model and a linear time delay mixing model; obtaining a hybrid matrix estimation value of the linear hybrid model through a fuzzy C-means clustering algorithm, and determining a hybrid matrix estimation error; performing source signal separation on the mixed signal through the linear mixed model and a norm minimum method to obtain a source signal estimated value; determining a cross correlation coefficient of the source signal estimation value and a source signal; and according to the hybrid matrix est