SAR automatic target recognition based on K-means and data augmentation

Synthetic aperture radar (SAR) automatic target recognition (ATR) has been receiving more and more attention in the past two decades. A lot of methods have been proposed and studied for radar target recognition. Among some of these methods, they use the supervised algorithms to extracts features. In this paper, we first use a unsupervised algorithm, K-means clustering, which can learn the features without known the class of training samples, for radar target recognition. As the unsupervised algorithm has a high demand on the scale of the data, so we proposed a method of data augmentation to get more data for the unsupervised algorithm, by which the K-means clustering algorithm can learn more unsupervised features. Experimental results on the MSTAR database show that the proposed method can achieve satisfying recognition accuracy compared with other state-of-the-art methods.