针对内河航迹数据存在结构复杂、密度差异显著及转向点分布不均等问题，本文采用基于层次密度的HDBSCAN聚类算法对典型水域的AIS轨迹数据进行了聚类分析。该方法具备自动识别簇数、适应多密度结构、剔除噪声点等优势，能够弥补传统聚类算法在处理不规则航迹簇时的局限性。通过构建多维特征向量并引入方向一致性约束，对聚类结果的空间分布与结构连贯性进行分析。结果表明，该方法可有效提取主通行轨迹及典型航迹段落，在不同密度区域均表现出良好的聚类稳定性与结构表达能力。研究成果可为轨迹结构建模与通航行为分析提供高质量的基础支撑。
To address the challenges posed by the complex structure, uneven density, and irregular turning points of inland vessel trajectories, this article applies the Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) algorithm to analyze Automatic Identification System (AIS) trajectory data in a typical inland waterway. The method is capable of automatically determining the number of clusters, adapting to varying density distributions, and identifying noise points, effectively overcoming limitations of traditional clustering techniques in handling irregular trajectory shapes. A multi-dimensional feature representation is constructed, incorporating directional consistency constraints to evaluate the spatial distribution and structural continuity of the clustering results. The analysis shows that the method can effectively extract dominant navigation trajectories and typical movement segments while maintaining structural coherence and clustering stability across diverse density regions. The results provide a reliable data foundation for subsequent modeling of trajectory structures and navigation behavior analysis.
2026,48(2): 145-150 收稿日期：2025-4-15
DOI：10.3404/j.issn.1672-7649.2026.02.023
分类号：U662.9
基金项目：重庆市自然科学基金项目（CSTB2025NSCQ-GPX0838）；重庆市教委科学技术研究项目（KJQN202500716）；重庆交通大学水利水运工程教育部重点实验室开放基金项目资助（SLK2023B15）
作者简介：何宣柳(2002-),男,硕士,研究方向为船舶路径规划及避碰
参考文献：
[1] HE Z, CHU X, LIU C, et al. A novel model predictive artificial potential field based ship motion planning method considering COLREGs for complex encounter scenarios[J]. ISA Transactions, 2023, 134: 58-73.
[2] GAO P, XU P, CHENG H, et al. Hybrid path planning for unmanned surface vehicles in inland rivers based on collision avoidance regulations[J]. Sensors, 2023, 23(19): 8326.
[3] 甄荣, 邵哲平, 潘家财. 基于AIS数据的船舶行为特征挖掘与预测: 研究进展与展望[J]. 地球信息科学学报, 2021, 23(12): 2111-2127.
ZHEN R, SHAO Z P, PAN J C. Advance in character mining and prediction of ship behavior based on AIS data[J]. Journal of Geo-information Science, 2021, 23(12): 2111-2127.
[4] 韩保隆. 基于深度学习和改进K-means文本聚类算法研究[D]. 大连: 大连交通大学, 2023.
[5] 赵小强, 姚青磊. 基于DBSCAN聚类分解和过采样的随机森林不平衡数据分类算法[J]. 兰州理工大学学报, 2023, 49(6): 80-89.
ZHAO X Q, YAO Q L. Classification algorithm of imbalanced data using DBSCAN decomposition and oversampling based random forest[J]. Journal of Lanzhou University of Technology, 2023, 49(6): 80-89.
[6] 周传华, 鲁勇, 于猜. 基于数据分区的OPTICS聚类算法[J]. 传感器与微系统, 2022, 41(10): 103-107.
ZHOU C H, LU Y, YU C. OPTICS clustering algorithm based on data partitioning[J]. Sensors and Microsystems, 2022, 41(10): 103-107.
[7] MCINNES L, HEALY J, ASTELS S. Hdbscan: hierarchical density based clustering[J]. Journal of Open Source Software, 2017, 2(11): 205.
[8] YANG J, LIU Y, MA L, et al. Maritime traffic flow clustering analysis by density based trajectory clustering with noise[J]. Ocean Engineering, 2022, 249: 111001.
[9] GAO M, SHI G Y. Ship spatiotemporal key feature point online extraction based on AIS multi-sensor data using an improved sliding window algorithm[J]. Sensors, 2019, 19(12): 2706.
[10] 郝永志. 基于AIS大数据的船舶路径规划方法研究[D]. 上海: 上海海事大学, 2023.
