Communication interference signal and spectrum hole joint cognitive method and device based on deep learning, and medium

The invention discloses a communication interference signal and spectrum hole joint cognition method and device based on deep learning and a medium. The method comprises the steps that a wireless signal data set with known interference types and whether spectrum holes exist or not is acquired from a signal receiver; wireless signals in the data set are processed to obtain baseband signals of the wireless signals, and power spectrums of the wireless signals are calculated; training a neural network based on residual connection and expansion convolution by taking the power spectrum of the signal as input and the corresponding interference type and the spectrum hole as output to obtain an interference and spectrum hole joint cognitive model; and for a newly received wireless signal of which the interference type and the spectrum hole are to be cognitive, obtaining the power spectrum according to the method, inputting the power spectrum into the interference and spectrum hole cognitive model, and outputting to ob