I. Introduction
Synthetic aperture radar (SAR) is an active observation system capable of continuously monitoring both land and coastal areas. It processes radar signals to generate SAR images, which have been applied in a variety of applications. Despite advances in SAR image technology, challenges remain in the detection of ships within SAR images [1], [2], specifically.
SAR image has complex background and variable environment, and the noise in the image is difficult to eliminate.
Ship targets in SAR image are densely arranged [3], which makes it challenging to distinguish between them and accurately outline their boundaries.
Ship targets vary greatly in scale in SAR images, posing a significant challenge to the adaptability of models.