Multiscale Ship Detection Method in SAR Images Based on Feature Enhancement

Ship detection from synthetic aperture radar (SAR) remote sensing images is essential for monitoring water traffic and marine safety. Numerous methods for ship detection have been developed. However, due to the scale diversity and the background clutter by the special imaging mechanism, ship detection in SAR images is still a substantial challenge. To tackle these issues, we propose a multi-scale ship detection method for SAR images based on attention mechanism and feature enhancement. A feature enhancement module (FEM) is embedded into a feature pyramid network (FPN). The FEM is introduced to reduce the influence of background clutter on feature extraction by integrating the center pixel difference convolution (CDC). In addition, considering the low-quality images, the WIoU is integrated. Finally, a spatial attention module is integrate to the detection head. Experiments on SSDD show that the proposed method achieves a comprehensive detection performance of 97.79% mean average precision (mAP) with the increase of 1.97%.