基于改进DBSCAN算法的受限航道船舶会遇规律挖掘研究

doi:10.3969/j.issn.1674-0696.2023.06.12

重庆交通大学学报（自然科学版） ›› 2023, Vol. 42 ›› Issue (6): 84-92.DOI: 10.3969/j.issn.1674-0696.2023.06.12

基于改进DBSCAN算法的受限航道船舶会遇规律挖掘研究

张新宇1, 姚沅辛1,齐壮1,黄力2,杨琼2

（1.大连海事大学航海学院，辽宁大连 116026；2.交通运输部规划研究院，北京，100028）

收稿日期:2021-01-15 修回日期:2023-05-29 出版日期:2023-07-07 发布日期:2023-08-01
作者简介:张新宇(1978—)，男，辽宁大连人，教授，主要从事海事大数据、船舶交通组织与调度、无人船方面的研究。E-mail：zhangxy@dlmu.edu.cn 通信作者：姚沅辛(1996—),女,吉林延边人,硕士，主要从事海事大数据方面的研究。E-mail：yyx96725@dlmu.edu.cn
基金资助:
国家自然科学基金项目(51779028)

Mining the Rules of Ships Encountering in Restricted Channel Based on Improved DBSCAN Algorithm

ZHANG Xinyu1, YAO Yuanxin1, QI Zhuang1, HUANG Li2, YANG Qiong2

(1.Navigation College, Dalian Maritime University, Dalian 116026, Liaoning, China; 2. Transport Planning and Research Institute, Ministry of Transport, Beijing 100028, China)

Received:2021-01-15 Revised:2023-05-29 Online:2023-07-07 Published:2023-08-01

摘要/Abstract

摘要： 受限航道船舶会遇态势复杂多变，为了识别潜在航行风险，提出一种基于SO-DBSCAN(self-optimal DBSCAN)算法的受限航道船舶会遇规律挖掘方法。首先，构建了基于船速比修正的受限航道动态船舶领域模型，考虑受限航道船舶安全间距以及航速对交通流的影响，引入船速比修正因子改进藤井模型；其次，提出一种基于动态舷角的会遇态势划分方法，结合国际海上避碰规则相关条款要求，采用船舶间动态舷角关系及航行速度作为引导划分出14种会遇态势；最后，引入Silhouette Coefficient和Calinski-Harabasz评估指标，并以加权和最大为目标函数自主求解最优的聚类策略，通过采集天津港受限航道的AIS数据进行实例验证，挖掘会遇时空规律并识别会遇热点。研究表明：该水域内船舶在15：00—17：00时间段内会遇较为频繁，会遇关键区域主要集中在42#~48#灯浮和22#~32#灯浮附近，结果符合航海实际；基于SO-DBSCAN算法的受限航道船舶会遇规律挖掘方法不仅能够解决传统基于密度聚类算法调参复杂的问题，还能够有效识别会遇高风险区域。

关键词: 交通+大数据人工智能, AIS数据, 受限航道, 船舶领域, 会遇态势, 航行安全

Abstract: The encounter situation of ships in restricted channel is complex and changeable. In order to identify the potential navigation risk, a method based on SO-DBSCAN (self-optimal DBSCAN) algorithm to mine the ship encounter laws in restricted channel was proposed. Firstly, a dynamic ship domain model of restricted channel based on ship speed ratio correction was constructed. Considering the influence of ship safety distance and navigation speed on traffic flow in restricted channel, the Fuji model was improved by introducing ship speed ratio correction factor. Secondly, a method of encounter situation division based on dynamic relative bearing was proposed. According to the requirements of the relevant provisions of the international regulations for preventing collisions at sea, 14 kinds of encounter situations were divided by using the dynamic relative bearing relationship between ships and the navigation speed as guidance. Finally, Silhouette Coefficient and Calinski-Harabasz evaluation indexes were introduced, and the maximum weighted sum was taken as the objective function to independently solve the optimal clustering strategy. By collecting the AIS data of the restricted channel of Tianjin port for example verification, the temporal and spatial rules of encountering were mined, and the hot spots of encountering were identified. The research shows that the encounter frequency of ships in the study area is more frequent in the period of 15:00—17:00, and the encounter critical waters are mainly concentrated in the vicinity of Buoy No.42-No.48 and Buoy No.22-No.32, which are in line with the actual navigation. In general, the proposed method to mine the ship encounter laws in restricted channel based on SO-DBSCAN algorithm can not only solve the problem of complex parameter adjustment based on traditional density clustering algorithm, but also effectively identify the high-risk areas.

Key words: traffic + big data artificial intelligence, AIS data, restricted channel, ship domain, encounter situation, navigation safety

中图分类号:

张新宇, 姚沅辛, 齐壮, 黄力, 杨琼. 基于改进DBSCAN算法的受限航道船舶会遇规律挖掘研究[J]. 重庆交通大学学报（自然科学版）, 2023, 42(6): 84-92.

ZHANG Xinyu, YAO Yuanxin, QI Zhuang, HUANG Li, YANG Qiong. Mining the Rules of Ships Encountering in Restricted Channel Based on Improved DBSCAN Algorithm[J]. Journal of Chongqing Jiaotong University(Natural Science), 2023, 42(6): 84-92.

参考文献

［1］张树奎,肖英杰,尤晓静.基于复杂网络的交叉航道内船舶汇聚度模型［J］.重庆交通大学学报(自然科学版),2017,36(2):95-100.
ZHANG Shukui, XIAO Yingjie, YOU Xiaojing. Model of ships aggregation at fairway crossing based on complex network ［J］. Journal of Chongqing Jiaotong University (Natural Science), 2017, 36(2): 95-100.
［2］ CHEN P F, HUANG Y M, PAPADIMITRIOU E, et al. An improved time discretized non-linear velocity obstacle method for multi-ship encounter detection ［J］. Ocean Engineering, 2020, 196:106718.
［3］ WESTRENEN F V, ELLERBROEK J. The effect of traffic complexity on the development of near misses on the North Sea ［J］. IEEE Transactions on Systems Man & Cybernetics Systems, 2017, 47(3): 432-440.
［4］ FUJII Y, TANAKA K. Traffic capacity ［J］. Journal of Navigation, 1971, 24(4): 543-552.
［5］ ZHANG W B, GOERLANDT F, KUJALA P, et al. An advanced method for detecting possible near miss ship collisions from AIS data ［J］. Ocean Engineering, 2016, 124:141-156.
［6］ CHAI T, WENG J X, LI G R. Estimation of vessel collision frequency in the Yangtze river estuary considering dynamic ship domains ［J］. Journal of Marine Science and Technology, 2020, 25(3): 964-977.
［7］李东枫. 基于AIS大数据的船舶危险会遇热点区域挖掘研究［D］. 广州：华南理工大学, 2017.
LI Dongfeng. Research on Dangerous Encounter Hotspots Mining Based on AIS Big Data ［D］. Guangzhou: South China University of Technology, 2017.
［8］甄荣, RIVEIRO M, 金永兴. 一种基于DBSCAN的船舶会遇实时识别方法［J］. 上海海事大学学报, 2018, 39(1):1-5.
ZHEN Rong, RIVEIRO M, JIN Yongxing. A real-time identification method to ship encounter based on DBSCAN ［J］. Journal of Shanghai Maritime University, 2018, 39(1):1-5.
［9］李永攀,刘正江,郑中义. 基于时空密度的船载AIS数据聚类分析方法研究［J］.重庆交通大学学报(自然科学版), 2018, 37(10): 117-122.
LI Yongpan, LIU Zhengjiang, ZHENG Zhongyi. Clustering analysis method of shipborne AIS data based on spatio-temporal density ［J］. Journal of Chongqing Jiaotong University (Natural Science), 2018, 37(10): 117-122.
［10］王炳. 基于AIS数据的水上应急救援基地选址研究［D］.广州：华南理工大学,2019.
WANG Bing. Research on Location Selection of Aquatic Emergency and Rescue Bases Based on AIS Data ［D］. Guangzhou: South China University of Technology, 2019.

基于改进DBSCAN算法的受限航道船舶会遇规律挖掘研究

Mining the Rules of Ships Encountering in Restricted Channel Based on Improved DBSCAN Algorithm

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 1

编辑推荐

Metrics