Classification of ground moving radar targets using convolutional neural network

Abstract

In this paper, we propose a new approach for Pulsed Doppler Radar Automatic Target Recognition (ATR). Target classification depends highly on the quality of the training database, the extracted features and the classification algorithm. Radar echo signals captured by the Radar show the Doppler effect produced by moving targets. Those echo signals can be processed in different domains to attain distinctive characteristics of targets that can be used for target classification. The proposed approach is based on utilizing a pre-trained Convolutional Neural Network (CNN) as a feature extractor whereas the output features are used to train a multiclass Support Vector Machine (SVM) classifier. Our approach was tested on RadEch database of 8 ground moving targets classes. Our approach outperformed the state-of-the-art methods, using the same database, and reached an accuracy of 99%.