舰船目标检测在海洋交通监控、军事海防预警等领域具有重要作用。本文围绕形态学重构在舰船目标分割中的优化展开研究，阐述形态学重构理论基础，分析舰船图像特征，构建“自适应初始化-约束迭代重构-精准优化修正”的分割模型。基于HRSID数据集与自建光学数据集开展实验，通过α/λ敏感性实验确定最优参数，复杂海况鲁棒性实验显示本文模型在10级强浪下交并比（IoU）达0.693，分割性能对比实验表明本文方法准确率89.8 %、F1分数88.5 %、单帧处理时间0.12 s。本文方法可应用于低标注数据、实时性要求高的舰船检测场景，为复杂海况下高精度舰船分割提供轻量化技术方案。
Ship target detection plays a significant role in fields such as Marine traffic monitoring and military coastal defense early warning. This paper focuses on the optimization of morphological reconstruction in ship target segmentation, elaborates the theoretical basis of morphological reconstruction, analyzes the characteristics of ship images, constructs a segmentation model of "adaptive initialization - constrained iterative reconstruction - precise optimization correction", conducts experiments based on the HRSID dataset and the self-built optical dataset, and determines the optimal parameters through α/λ sensitivity experiments. The robustness experiments of complex sea conditions show that the intersection and union ratio (IoU) of the model proposed in this paper reaches 0.693 under a strong wave of level 10. The comparison experiments of segmentation performance indicate that the accuracy rate of the method proposed in this paper is 89.8%, the F1 score is 88.5%, and the single-frame processing time is 0.12 s. The method proposed in this paper can be applied to the ship detection scenarios with low labeled data and high real-time requirements, providing a lightweight technical solution for high-precision ship segmentation in complex sea conditions.
2025,47(22): 175-179 收稿日期：2025-5-11
DOI：10.3404/j.issn.1672-7649.2025.22.026
分类号：U667.3；TP391.41
作者简介：陆军（1980 – ），男，硕士，讲师，研究方向为计算机辅助工业设计及人机工程、舰船目标识别
参考文献：
[1] 徐新瑶, 王小龙. 基于Radon变换的高分辨SAR图像舰船目标精细分割[J]. 电子技术应用, 2023, 49(5): 142-148
XU X Y, WANG X L. Fine segmentation of ship targets in high-resolution sar images based on radon transform[J]. Application of Electronic Technology, 2023, 49(5): 142-148
[2] 李震宇, 高兵, 郭德明, 等. 强海杂波下机载雷达HRRP舰船长度提取算法[J]. 现代雷达, 2022, 44(04): 23-30
LI Z Y, GAO B, GUO D M, et al. Ship length extraction algorithm for airborne radar HRRP under strong sea clutter[J]. Modern Radar, 2022, 44(4): 23-30
[3] 丁鹏, 张叶, 贾平, 等. 基于多尺度多特征视觉显著性的海面舰船检测[J]. 光学精密工程, 2017, 25(9): 2461-2468
DING P, ZHANG Y, JIA P, et al. Sea surface ship detection based on multi-scale and multi-feature visual saliency[J]. Optics and Precision Engineering, 2017, 25(9): 2461-2468
[4] 郭政, 赵梅, 胡长青, 等. 舰船辐射噪声广义多尺度数学形态学特征提取与应用研究[J]. 振动与冲击, 2022, 41(4): 21-28+100.
GUO Z, ZHAO M, HU C Q, et al. Research on generalized multi-scale mathematical morphology feature extraction and application of ship radiated noise[J]. Journal of Vibration and Shock, 2022, 41(4): 21-28+100.
[5] 罗兴潮, 黄文骞, 程益锋. 大范围近海舰船近红外遥感检测方法[J]. 舰船电子工程, 2019, 39(10): 216-219+236
LUO X C, HUANG W Q, CHENG Y F. Near-infrared remote sensing detection method for ships in large-scale coastal waters[J]. Ship Electronic Engineering, 2019, 39(10): 216-219+236
