基于双边频域空间域特征增强的水面目标检测算法

iming at the challenges in water surface object detection, including wave interference, target occlusion, and the difficulty of detecting distant small objects, we propose an improved YOLOv11-based detection algorithm. Specifically, to address issues such as blurred boundaries of ships, birds, buoys, and waves, we employ Fourier transform in the frequency domain to enhance the capture of edge, texture, and structural features, improving recognition in complex backgrounds. Additionally, the SwinTransformer architecture is integrated to capture long-range dependencies and strengthen global feature extraction. A shallow detector is designed to optimize the detection of small objects and improve overall performance. Experimental results demonstrate that this approach significantly enhances the robustness and accuracy of water surface object detection, providing strong practical value for large-scale maritime surveillance and vessel monitoring.