传统依赖人工目视或摄像设备的海上目标识别方式早已无法满足现代海洋监测的需求，采用智能识别技术成为当前提升海上安全与资源管理效率的关键手段。为满足海上目标实时的高精度识别需求，提出一种基于分层推理的海上目标智能识别方法。通过构建YOLOv8、ResNet50联合分层推理模型，采用粗检测、细分类逻辑嵌套、逐步递进的方式提高识别速度与准确率。基于建立的海上船舶目标数据集开展实验，结果表明，本文提出的分层推理模型检测准确率达到90.2%，每秒帧数73.0，优于其他单层图像检测模型，且能够较好地平衡海上目标图像识别的精度和效率，实现海上目标在细分类下的准确、快速识别。
Traditional maritime target identification methods relying on manual visual inspection or camera devices can no longer meet modern marine monitoring requirements. Adopting intelligent recognition technology has become a key solution for enhancing maritime safety and resource management efficiency. To meet the demand for real-time and high-precision identification of maritime targets, this paper proposes an intelligent identification method for maritime targets based on hierarchical reasoning. By constructing a hierarchical reasoning model combining YOLOv8 and ResNet50, the study employs a coarse classification and fine recognition logic nested approach to progressively improve identification speed and accuracy. Experiments conducted on the established maritime vessel target dataset demonstrate that our proposed hierarchical inference model achieves 90.2% detection accuracy at 73.0 FPS, outperforming other single-layer image detection models. It effectively balances the precision and efficiency of maritime target image recognition, achieving accurate and rapid identification under fine classification.
2025,47(24): 191-195 收稿日期：2025-5-15
DOI：10.3404/j.issn.1672-7649.2025.24.031
分类号：U675.7；TP391.4
作者简介：许莹琪(1997-),女,硕士,助理工程师,研究方向为智能算法
参考文献：
[1] 马啸, 邵利民, 金鑫, 等. 改进的YOLO模型及其在舰船目标识别中的应用[J]. 电讯技术, 2019, 59(8): 869-874.
[2] LECUN Y, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.
[3] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[C]//Proceedings of the 25th International Conference on Neural Information Processing Systems, 2012.
[4] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[J]. Computer Science, 2014: 1409-1556.
[5] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[J]. IEEE, 2016.
[6] LIU Z, MAO H, WU C Y, et al. A ConvNet for the 2020s[J]. Arxiv E-prints, 2022.
[7] REN S, HE K, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2017, 39(6): 1137-1149.
[8] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016.
[9] WEI L, DRAGOMIR A, DUMITRU E, et al. SSD: Single shot multibox detector[J]. Springer, Cham, 2016.
[10] 文莉莉, 孙苗, 邬满. 基于Faster R-CNN的海域监管预警方法[J]. 吉林大学学报(信息科学版), 2021, 39(4): 421-429.
[11] 张晓鹏, 许志远, 曲胜, 等. 基于改进YOLOv5深度学习的海上船舶识别算法[J]. 大连海洋大学学报, 2022, 37(5): 866-872.
[12] 张建东. 融合深度监督与改进YOLOv8的海上目标检测[J]. 南京信息工程大学学报, 2024, 16(4): 482-489.
[13] EZZEDDINI L, KTARI J, FRIKHA T, et al. Analysis of the performance of Faster R-CNN and YOLOv8 in detecting fishing vessels and fishes in real time [J]. Computer Science, 2019: 9246-9255.
[14] 苏睿涵, 胡剑秋, 蔡庆, 等. 面向气象无人艇的实时海面场景分类方法[J]. 舰船科学技术, 2025, 47(6): 88-93.
SU R H, HU J Q, CAI Q, et al. Real-time sea surface scene classification method for meteorological unmanned surface vehicles[J]. Ship Science and Technology, 2025, 47(6): 88-93.
[15] 赵圆圆, 李月军, 李昌庆, 等. 基于深度神经网络的船舶图像识别检索研究[J]. 舰船科学技术, 2024, 46(3): 174-177.
ZHAO Y Y, LI Y J, LI C Q, et al. Research on ship image recognition and retrieval based on deep neural network[J]. Ship Science and Technology, 2024, 46(3): 174-177.
