Drones and helicopters classification using point clouds features from radar

This paper addresses the issue of drones and helicopters classification from radar return using point clouds features. In order to classify helicopters or drones, most published researches estimate the length of blades, the number of rotors, the period of their micro-Doppler signatures or try to extract information from spectrograms (or any other image). Differently this work generates a point cloud from radar return signal. Features are extracted from point cloud and they are used as input of an Artificial Neural Net (ANN) to classify different models of helicopters and drones.