Abstract

Different from optical image limited in time and space, sparse synthetic aperture radar (SAR) can obtain high-resolution image in all day and all weather conditions. Thus, SAR image has been widely used in military fields. With rapid development of technology and the growth of data volume, it is obviously impractical to interpret SAR image manually. Therefore, in this paper, we propose a novel sparse SAR automatic target recognition (ATR) framework, which is composed by the regularization sparse recovery algorithm and YOLO networks, so as to identify the SAR target quickly and accurately. In the proposed framework, we first use the complex approximate message passing (CAMP) based sparse image recovery algorithm to construct the sparse SAR image dataset, then identify the SAR targets by YOLOv3 and YOLOv4 networks. Experimental results based on MSTAR dataset validate the proposed framework effectively.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call