Arbitrary-Direction SAR Ship Detection Method for Multiscale Imbalance

Zhongzhen Sun, Xiangguang Leng, Xianghui Zhang, Zheng Zhou, Boli Xiong, Kefeng Ji, Gangyao Kuang

Abstract: The arbitrary-oriented ship detection in synthetic aperture radar (SAR) imagery remains especially challenging due to multiscale imbalance and the characteristics of SAR imaging, a problem that is more pronounced than in optical ship detection. Unlike optical images, SAR data often lack rich textural and color cues, instead exhibiting nonuniform scattering, speckle noise, and nonstandard elliptical ship shapes, all of which make robust feature extraction and bounding box regression significantly more difficult across different scales. To address these unique SAR-specific challenges, this article proposes the multiscale dynamic feature fusion network (MSDFF-Net) aims to alleviate multiscale imbalance in three main ways. First, a multiscale large-kernel convolutional block (MSLK-Block) integrates large-kernel convolutions with partitioned heterogeneous operations to enhance the multiscale feature representation, tackling wide-ranging ship sizes under noisy conditions. Second, a dynamic feature fusion block (DFF-Block) handles scale-based feature utilization imbalance by adaptively balancing spatial and channel information, thereby reducing interference from clutter and strengthening discrimination for diverse-scale ships. Third, we propose the Gaussian probability distribution (GPD) loss function, which models ships’ elliptical scattering properties and mitigates regression loss imbalance for targets of varying scales and orientations. Experimental evaluations on the R-SSDD, R-HRSID, and CEMEE datasets demonstrate that MSDFF-Net reaches top-tier performance standards, outperforming 21 existing deep learning (DL)-based SAR ship detectors. Specifically, MSDFF-Net achieves 93.95% precision, 94.72% recall, 91.55% mean average precision (mAP), 94.33% $F1$ -score, and 135.79 frames/s (FPS) on the R-SSDD dataset, with a parameter size of only 8.94 M. In addition, MSDFF-Net exhibits strong transferability across large-scale SAR images, making it suitable for real-world deployment. The code and datasets can be accessed publicly at: https://github.com/SZZ-SXM/MSDFF-Net

External IDs:dblp:journals/tgrs/SunLZZXJK25