Unmanned aerial vehicle time-frequency atlas database construction method based on frequency domain data enhancement and expansion

The invention discloses an unmanned aerial vehicle time-frequency atlas database construction method based on frequency domain data enhancement and expansion. The method comprises the following implementation steps: acquiring an unmanned aerial vehicle signal; constructing a time-frequency matrix; constructing an initial time-frequency matrix database; expanding the time-frequency database by using a frequency domain data enhancement method; training a residual network ResNet by using a time-frequency matrix in the initial database; and testing the ResNet by using the expanded database. The problems that the application range of collected data is narrow, and an unmanned aerial vehicle time-frequency spectrum sample neural network cannot be trained under the condition that an unmanned aerial vehicle time-frequency spectrum database constructed under the limitation of a specific task lacks partial channels are solved. The database constructed by the method has the advantages that samples are sufficient and can