Radar Target Classification Using Deep Learning

Abstract

This chapter describes various applications of deep learning algorithms to radar images. In particular, the classification of micro-Doppler spectrograms, range-Doppler diagrams, and synthetic aperture radar images is addressed in terms of convolutional neural networks and recurrent neural networks. The applications discussed include human motion classification, hand gesture recognition, drone detection, vehicle detection, ship detection, and more. Advanced deep learning techniques, such as transfer learning, generative adversarial networks, and continual learning, are also applied to radar images, and their performance is evaluated.