准确识别海上目标对提高舰船航行安全、维护海上权益意义重大。YOLOv7作为YOLO系列算法的最新成果，在目标检测任务中拥有良好的速度和精度。但通用化的网络应用于特定场景时，由于权重过大可移植性差，优势并不明显。本文根据海上目标分布及背景特点，利用注意力机制提升网络的特征提取与特征融合能力，提出一种基于注意力融合的海上目标检测算法CS-YOLOv7s。在新加坡海上数据集中的实验结果表明，本文所提网络在少量降低准确率的同时，大幅度降低网络权重，提高检测速度，可满足海上目标实时检测任务需要。
Accurate identification of maritime targets is of great significance to improving the navigation safety of ships and safeguarding my country’s maritime rights and interests. As the latest achievement of YOLO series algorithms, YOLOv7 has good speed and accuracy in target detection tasks. However, when the generalized network is applied in a specific scenario, the advantage is not obvious due to the large weight and poor portability. According to the distribution and background characteristics of marine targets, the attention mechanism is used to improve the feature extraction and feature fusion capabilities of the network, and a marine target detection algorithm CS-YOLOv7s based on attention fusion is proposed. The experimental results in the Singapore maritime data set show that the proposed network can greatly reduce the network weight and improve the detection speed while reducing the accuracy slightly, which can meet the needs of real-time detection tasks of maritime targets.
2023,45(16): 120-124 收稿日期：2022-7-21
DOI：10.3404/j.issn.1672-7649.2023.16.024
分类号：TP391
基金项目：国家自然科学基金资助项目（61501486）
作者简介：张鹏(1994-),男,硕士研究生,研究方向为系统工程
参考文献：
[1] 薛董敏. 基于视觉显著性的海上远景目标检测系统[J]. 舰船科学技术, 2021, 43(24): 184–186
[2] 包晓敏, 王思琪. 基于深度学习的目标检测算法综述[J]. 传感器与微系统, 2022, 41(4): 5–9
[3] 赵文思. YOLOv3在船舶图像目标检测应用研究[D].大连: 大连理工大学, 2021.
[4] 王文亮, 李延祥, 张一帆, 等. MPANet-YOLOv5: 多路径聚合网络复杂海域目标检测[J]. 湖南大学学报(自然科学版), 2022, 49(10): 69-76.
[5] 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.
[6] 姬嗣愚, 王永生. 基于改进YOLOV5的海上弹着点水柱信号检测算法[J]. 电光与控制, 2022, 29(8): 50-56.
[7] WANG C Y, BOCHKOVSKIY A, LIAO H Y M. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[J]. arXiv preprint arXiv, 2022.
[8] NEUBECK A, GOOL L. Efficient non-maximum suppression[C]//International Conference on Pattern Recognition. IEEE Computer Society, 2006.
[9] SOUDY M, AFIFY Y, BADR N. RepConv: a novel architecture for image scene classification on Intel scenes dataset[J]. International Journal of Intelligent Computing and Information Sciences, 2022, 22(2): 63–73
[10] LEE Chen-Yu, XIE Saining, GALLAGHER Patrick, et al. Deeply-supervised nets[J]. In Artificial Intelligence and Statistics, 2015(5): 562−570.
[11] MEHTA S, RASTEGARI M. Mobilevit: light-weight, general-purpose, and mobile-friendly vision transformer[J]. arXiv preprint arXiv: 2110.02178, 2021.
[12] WOO S, PARK J, LEE J Y, et al. Cbam: convolutional block attention module[C]//Proceedings of the European conference on computer vision (ECCV), 2018: 3–19.
