基于改进YOLOv11的海洋内波目标检测算法OA

Aiming at the problem of insufficient about internal wave feature extraction and lack of adaptability of feature fusion in complex marine environment,this paper proposes a ocean internal wave target detection algorithm MSBA-YOLO based on improved YOLOv11 model.Firstly,a hierarchical scale feature aggregation module is designed to effectively extract the spatial and texture features of internal waves.Secondly,a dual attention fusion network is constructed to realize adaptive adjustment and weighted fusion of different scale features.Finally,the statistical token attention module is introduced to accurately locate the internal wave target while reducing the computational complexity.The experimental verification based on Sentinel-1 SAR image dataset shows that the mAP50 of MSBA-YOLO reaches 82.4%,which is 4.3%higher than the baseline model,and the number of parameters is only 2.7 M.The research results show that the algorithm significantly improves the positioning ability of internal wave characteristics under complex sea conditions while maintaining very low computational overhead and high real-time performance,and provides efficient and accurate technical support for marine engineering safety early warning and internal wave dynamic monitoring.