Unmanned Aerial Vehicles Identified Based on Transfer Learning of Deep Convolutional Neural Network

This paper proposes a method based on transfer learning to identify the Unmanned Aerial Vehicles (UAVs). This method uses the UAVs echoes to construct a micro-Doppler spectrogram (MDS) dataset by the time-frequency analysis. The Deep Convolutional Neural Network (DCNN) is trained by the MDS dataset. The transfer learning is used to improve the UAVs identification accuracy of the DCNN.