An End-to-End Network for Multi-Scale Ship Detection in SAR Images
Hongwen Dong, Feiming Wei, Qianqian Xu, Xiao Wang, Gao Sun · 2023
Automatic target recognition (ATR) technology of synthetic aperture radar (SAR) image is one of the key technologies of artificial image interpretation. Most of the traditional SAR ship target detection algorithms are limited by the scene and have poor generalization ability. In this paper, an end-to-end network for multi-scale ship detection in SAR images is proposed, which combines a multi-scale feature extraction module and an attention module to solve these problems. The multi-scale feature extraction module enhances the representation ability of the model to target size variations ad appearance changes by exploiting context information of different resolutions. The attention module enhances the target information and suppresses the background redundancy information in the spatial and channel dimensions of the feature. In the experiment, the effectiveness of the proposed method is verified on the public SAR ship detection dataset (SSDD), and the proposed method achieves 0.998 average precision (AP).