Automatic recognition of ISAR images based on deep learning

The problem of target classification of non-cooperative airplane using inverse synthetic aperture radar (ISAR) images is studied. ISAR image's variation with the radar imaging view makes the classification difficult. The point scatter models of five different airplanes are built. Deep learning methods have been shown to be able to classify images in a variety of disciplines with a high level of accuracy. As an effective deep learning algorithm, sparse autoencoder (SA) learning algorithm can automatically learn features from unlabeled data. In this paper, the SA algorithm is used to extract the features of the ISAR images. The extracted features are served as the input of the softmax classifier and then the recognition and classification of ISAR images are achieved. Simulation results verify the effectiveness and superiority of the proposed method.