Ship trajectories clustering method considering similarity in geometric position and mobility features

doi:10.11947/j.AGCS.2025.20240384

Abstract

Abstract:

With the widespread application of the automatic identification system (AIS) in maritime management, a large amount of vessel trajectory data is being recorded. By applying clustering analysis to this data, typical navigation routes and maritime area distribution information of vessels can be effectively mined, providing data support for shipping safety. However, existing clustering methods tend to focus on the spatial shape characteristics of vessel trajectories, and while some studies have addressed the movement characteristics of vessels, the exploration of deep-level information remains insufficient. This paper proposes a vessel trajectory clustering method that comprehensively considers both geometric position and movement feature similarity. Firstly, a trajectory classification model based on a one-dimensional convolutional neural network (U-net) is trained, and the feature vectors obtained before the model predicts the vessel type are used as deep-level features representing the vessel's movement pattern. Next, the Euclidean distance is employed to measure the similarity between these deep-level features, quantifying the similarity of vessel trajectories in terms of movement feature. Meanwhile, the Hausdorff distance is used to measure the geometric position similarity between different vessel trajectories. Subsequently, the fusion distance, combining geometric position and movement feature similarity, is calculated. Based on this distance, an adaptive density-based spatial clustering of applications with noise (DBSCAN) algorithm is applied for clustering analysis. Experimental results show that the proposed clustering method achieves superior clustering performance compared to baseline models in complex vessel trajectory application scenarios, effectively classifying trajectories with similar geometric shapes but different motion patterns into different clusters, thus providing more granular clustering results.

Key words: ship trajectory clustering, geometric position similarity, mobility feature similarity, deep learning

CLC Number:

P208

Xiaoya AN, Weiru GUO, Pengxin ZHANG, Xinxin LI, Lei SHI. Ship trajectories clustering method considering similarity in geometric position and mobility features[J]. Acta Geodaetica et Cartographica Sinica, 2025, 54(6): 1107-1121.

Figures/Tables 16

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Tab. 1

Fig. 7

Fig. 8

Fig. 9

Fig. 10

Tab. 2

Fig. 11

Tab. 3

Fig. 12

Fig. 13

References 29

[1]	刘海砚, 郭漩, 刘俊楠. 时空和语义结合的船舶轨迹数据压缩方法[J]. 测绘学报, 2023, 52(11): 1974-1982. DOI: . doi: 10.11947/j.AGCS.2023.20210658
	LIU Haiyan, GUO Xuan, LIU Junnan. A vessel trajectory data compression method combining spatio-temporal and semantic features[J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(11): 1974-1982. DOI: . doi: 10.11947/j.AGCS.2023.20210658
[2]	GUHA S, RASTOGI R, SHIM K. Cure: an efficient clustering algorithm for large databases[J]. Information Systems, 2001, 26(1): 35-58.
[3]	PALLOTTA G, VESPE M, BRYAN K. Vessel pattern knowledge discovery from AIS data: a framework for anomaly detection and route prediction[J]. Entropy, 2013, 15(6): 2218-2245.
[4]	ZHANG Rui, WU Hanyue, YIN Zhenzhong, et al. Predictive clustering of vessel behavior based on hierarchical trajectory representation[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(12): 19496-19506.
[5]	JIANG Yan, LI Bo, ZHANG Hao, et al. A novel classification scheme of moving targets at sea based on ward's and K-means clustering[C]//Proceedings of the 2nd International Conference on Computer Science and Application Engineering. Hohhot: ACM Press, 2018: 1-5.
[6]	ZHEN Rong, JIN Yongxing, HU Qinyou, et al. Maritime anomaly detection within coastal waters based on vessel trajectory clustering and Naïve Bayes classifier[J]. Journal of Navigation, 2017, 70(3): 648-670.
[7]	MA Wenyao, WU Zhaolin, YANG Jiaxuan, et al. Vessel motion pattern recognition based on one-way distance and spectral clustering algorithm[M]//Algorithms and Architectures for Parallel Processing. Cham: Springer International Publishing, 2014: 461-469.
[8]	GAO Miao, SHI Guoyou. Ship-handling behavior pattern recognition using AIS sub-trajectory clustering analysis based on the T-SNE and spectral clustering algorithms[J]. Ocean Engineering, 2020, 205: 106919.
[9]	ZHAO Liangbin, SHI Guoyou, YANG Jiaxuan. An adaptive hierarchical clustering method for ship trajectory data based on DBSCAN algorithm[C]//Proceedings of 2017 IEEE International Conference on Big Data Analysis. Beijing: IEEE, 2017: 329-336.
[10]	ZHAO Liangbin, SHI Guoyou. A trajectory clustering method based on Douglas-Peucker compression and density for marine traffic pattern recognition[J]. Ocean Engineering, 2019, 172: 456-467.
[11]	LI Ye, REN Hongxiang. Visual analysis of vessel behaviour based on trajectory data: a case study of the Yangtze River Estuary[J]. ISPRS International Journal of Geo-Information, 2022, 11(4): 244.
[12]	WANG Lianhui, CHEN Pengfei, CHEN Linying, et al. Ship AIS trajectory clustering: an HDBSCAN-based approach[J]. Journal of Marine Science and Engineering, 2021, 9(6): 566.
[13]	甄荣, 石自强. 一种基于高斯混合模型的船舶航迹聚类方法[J]. 船舶工程, 2021, 43(11): 139-143.
	ZHEN Rong, SHI Ziqiang. Ship trajectory clustering method based on Gaussian mixture model[J]. Ship Engineering, 2021, 43(11): 139-143.
[14]	YAO Di, ZHANG Chao, ZHU Zhihua, et al. Learning deep representation for trajectory clustering[J]. Expert Systems, 2018, 35(2): e12252.
[15]	LIANG Maohan, LIU R W, LI Shichen, et al. An unsupervised learning method with convolutional auto-encoder for vessel trajectory similarity computation[J]. Ocean Engineering, 2021, 225: 108803.
[16]	WANG Chao, HUANG Jiahui, WANG Yongheng, et al. A deep spatiotemporal trajectory representation learning framework for clustering[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(7): 7687-7700.
[17]	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 Rong, SHAO Zheping, PAN Jiacai. Advance in character mining and prediction of ship behavior based on AIS data[J]. Journal of Geo-Information Science, 2021, 23(12): 2111-2127.
[18]	ZHONG Hanyang, SONG Xin, YAN Zhenguo. Vessel sailing patterns analysis from S-AIS data based on K-means clustering algorithm[C]//Proceedings of the 4th International Conference on Big Data Analytics. Suzhou: IEEE, 2019: 10-13.
[19]	LIU Lei, ZHANG Yong, HU Yue, et al. A hybrid-clustering model of ship trajectories for maritime traffic patterns analysis in port area[J]. Journal of Marine Science and Engineering, 2022, 10(3): 342.
[20]	安健鹏, 李海霞, 雷亚丽, 等. 基于大数据的船舶活动轨迹规律挖掘方法[J]. 火力与指挥控制, 2024, 49(4): 156-163.
	AN Jianpeng, LI Haixia, LEI Yali, et al. Mining method of ship activity trajectory pattern based on big data[J]. Fire Control & Command Control, 2024, 49(4): 156-163.
[21]	ZHANG Yuanqiang, MA Yong, LIU Jiao. Ship trajectory segmentation and semisupervised clustering via geospatial background knowledge[J]. Ocean Engineering, 2024, 304: 117872.
[22]	刘海杨, 孟令航, 林仲航, 等. 基于轨迹点聚类的航路发现方法[J]. 计算机应用, 2022, 42(3): 890-894.
	LIU Haiyang, MENG Linghang, LIN Zhonghang, et al. Route discovery method based on trajectory point clustering[J]. Journal of Computer Applications, 2022, 42(3): 890-894.
[23]	张鹏鑫, 李连营, 杨敏, 等. 顾及运动行为特征变化的船舶轨迹分类模型[J]. 测绘科学, 2023, 48(5): 25-34.
	ZHANG Pengxin, LI Lianying, YANG Min, et al. A ship trajectory classification model considering changes in characteristics of motion behavior[J]. Science of Surveying and Mapping, 2023, 48(5): 25-34.
[24]	RONNEBERGER O, FISCHER P, BROX T. U-Net: convolutional networks for biomedical image segmentation[M]//Medical Image Computing and Computer-Assisted Intervention-MICCAI. Cham: Springer International Publishing, 2015: 234-241.
[25]	李文杰, 闫世强, 蒋莹, 等. 自适应确定DBSCAN算法参数的算法研究[J]. 计算机工程与应用, 2019, 55(5): 1-7, 148.
	LI Wenjie, YAN Shiqiang, JIANG Ying, et al. Research on method of self-adaptive determination of DBSCAN algorithm parameters[J]. Computer Engineering and Applications, 2019, 55(5): 1-7, 148.
[26]	WANG Lianhui, CHEN Pengfei, CHEN Linying, et al. Ship AIS trajectory clustering: an HDBSCAN-based approach[J]. Journal of Marine Science and Engineering, 2021, 9(6): 566.
[27]	万佳, 胡大裟, 蒋玉明. 多密度自适应确定DBSCAN算法参数的算法研究[J]. 计算机工程与应用, 2022, 58(2): 78-85.
	WAN Jia, HU Dasha, JIANG Yuming. Research on method of multi-density self-adaptive determination of DBSCAN algorithm parameters[J]. Computer Engineering and Applications, 2022, 58(2): 78-85.
[28]	ROUSSEEUW P J. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis[J]. Journal of Computational and Applied Mathematics, 1987, 20: 53-65.
[29]	DAVIES D L, BOULDIN D W. A cluster separation measure[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1979, 1(2): 224-227.

预测类型	真实类型	拖船	渔船	帆船/游艇	客船	货船	精确率	召回率	F₁值
一维U-net	拖船	160	7	2	1	8	0.94	0.90	0.92
	渔船	1	166	28	0	1	0.89	0.85	0.87
	帆船/游艇	2	13	196	6	2	0.84	0.90	0.87
	客船	0	0	6	176	2	0.92	0.96	0.94
	货船	7	0	0	8	170	0.93	0.92	0.92
LSTM	拖船	136	6	14	3	19	0.87	0.76	0.81
	渔船	5	151	21	9	10	0.87	0.77	0.82
	帆船/游艇	8	15	169	19	8	0.76	0.77	0.77
	客船	1	1	15	160	7	0.72	0.87	0.79
	货船	6	0	2	32	145	0.77	0.78	0.78

船舶轨迹聚类方法	轮廓系数	DBI指数	噪声率/（%）
基于Hausdorff距离的密度聚类方法	0.811	0.927	3.09
基于LCSS距离的密度聚类方法	0.801	0.915	4.45
MD-DBSCAN	0.792	0.965	7.45
顾及几何位置和移动特征相似性的聚类方法	0.815	0.876	4.63

船舶轨迹聚类方法	相似性度量		自适应确定DBSCAN聚类参数	聚类簇数	噪声率/（%）
船舶轨迹聚类方法	几何位置	移动特征	自适应确定DBSCAN聚类参数	聚类簇数	噪声率/（%）
基于Hausdorff距离度量几何位置相似性的聚类方法	k₁=1	k₂=0	ε=0.069	10	2.1
基于Hausdorff距离度量几何位置相似性的聚类方法	k₁=1	k₂=0	MinLns=5	10	2.1
顾及几何位置和移动特征相似性的聚类方法	k₁=0.92	k₂=0.08	ε=0.059	26	6.2
顾及几何位置和移动特征相似性的聚类方法	k₁=0.92	k₂=0.08	MinLns=3	26	6.2