Skip to Main Content
In this paper we propose an automatic target recognitionalgorithm to recognize and distinguish of three classes oftargets: personnel, wheeled vehicles and animals, based on thedoppler signatures they produce when moving, using a low-resolutionground surveillance pulse doppler RADAR. Using signal processingtechniques such as the short-time Fourier transform (STFT)and the chirplet transform, various parameters of the targets canbe extracted from the signal. This paper proposed classification ofmoving targets with use of the chirplet transform. The algorithmwas trained and tested on real radar signatures of multiple examplesof moving targets from each class and the performance wasshown to be invariant to target speed and orientation.
Date of Conference: 1-3 April 2008