地球信息科学学报 ›› 2018, Vol. 20 ›› Issue (5): 640-646.doi: 10.12082/dqxxkx.2018.180053

• “海上丝绸之路空间数据分析”专辑 • 上一篇    下一篇

上海外高桥港区停泊船聚类分析与异常检测

郑海林1,2(), 胡勤友1,*(), 杨春1, 陈金海3,4, 梅强3,4   

  1. 1. 上海海事大学 商船学院, 上海 201306
    2. 浙江海洋大学 港航与交通运输工程学院, 舟山 316022
    3.集美大学 航海学院, 厦门 361021
    4. 船舶辅助导航技术国家地方联合工程研究中心, 厦门 361021
  • 收稿日期:2018-01-05 修回日期:2018-03-22 出版日期:2018-05-29 发布日期:2018-05-20
  • 通讯作者: 胡勤友 E-mail:hlzhzjou@126.com;qyhu@shmtu.edu.cn
  • 作者简介:

    作者简介:郑海林(1987-),男,博士生,讲师,主要从事海事信息处理研究。E-mail: hlzhzjou@126.com

  • 基金资助:
    上海市科学技术委员会项目(15590501600);国家自然科学基金项目(41501490);中国科学院重点部署项目(ZDRW-ZS-2016-6-3);福建省教育厅基金项目(B16095)

Clustering Analysis and Anomaly Detection of Berthing Ships at Waigaoqiao Harbour District of Shanghai

ZHENG Hailin1,2(), HU Qinyou1,*(), YANG Chun1, CHEN Jinhai3,4, MEI Qiang3,4   

  1. 1. Merchant Marine College, Shanghai Maritime University, Shanghai 201306, China
    2. School of Port and Transportation Engineering, Zhejiang Ocean University, Zhoushan 316022, China
    3. Navigation College, Jimei University, Xiamen 361021,China
    4. National-local Joint Engineering Research Center for Marine Navigation Aids Services, Xiamen 36102, China
  • Received:2018-01-05 Revised:2018-03-22 Online:2018-05-29 Published:2018-05-20
  • Contact: HU Qinyou E-mail:hlzhzjou@126.com;qyhu@shmtu.edu.cn
  • Supported by:
    Project of Shanghai Municipal Science and Technology Commission, No.15590501600;National Natural Science Foundation of China, No.41501490;Key Project of the Chinese Academy of Sciences, No.ZDRW-ZS-2016-6-3;Fujian Provincial Educational Department Foundation, No.B16095.

摘要:

停泊船空间分布规律挖掘,在海事监管、港口管理和航运公司船队管理方面有着重要意义。现有研究主要针对船舶停泊点进行空间聚类以识别码头和锚地,缺乏对码头、锚地内船舶停泊特征分析,及码头和锚地外的异常停船的检测。因此,利用海量船舶自动识别系统(AIS)数据探索船舶停泊规律显得很有必要,且具备可行性。根据海况设定停泊速度阈值和停泊位置变化量阈值,建立停船判定模型。按港区、船型筛选,获取2016年1至11月外高桥港区集装箱船停泊记录。根据类中心点密度和聚类数量,设定邻域半径(ε)和邻域密度(MinPts),采用密度聚类(DBSCAN)算法对船舶停泊点进行密度聚类,并将聚类结果与外高桥港区码头、锚地分布图进行比较,生成可疑停船列表。对比船舶历史轨迹,明确可疑停船列表中船舶真实停泊记录,筛选出异常停船。研究发现,2016年1至11月外高桥港区船舶异常停泊点位于圆圆沙锚地至吴淞口锚地间的南港水道和江亚南沙锚地附近的南港水道航段。船舶停泊前、后位置变化幅度小,而速度变化幅度大,推测船舶突发故障是其异常停泊的原因。海事主管部门(MSA)可根据船舶水上移动通信业务识别码(MMSI)快速锁定航运公司,加强岸上船舶安全管理。船舶停泊位置和时间能够记录船舶发生故障地点及其持续时间,为船队管理提供重要依据。

关键词: 海上丝绸之路, 船舶自动识别系统, 停泊船, 密度聚类, 异常检测

Abstract:

Mining of the spatial distribution of berthing ships is of great significance in the maritime supervision, port management and fleet management of a shipping company. However, almost all of the existing studies focused on the spatial clustering of ship berthing points to identify berths and anchorages, and only few articles focused on the analysis of berthing ships' features in ports and the detection of the anomalous berthing ships outside the berths and anchorages. Therefore, it is necessary to use the massive automatic identification system (AIS) data to acquire the ship berthing features, which is also feasible due to general equipment of AIS on ships. By setting the threshold value of berthing speed and Variation of the berthing position according to the sea conditions, the model for determining the berthing ships could be established. Filtering by port area and ship type, we could obtain the container ship berthing records at Waigaoqiao Harbour District from January to November 2016. With the purpose of obtaining density distribution of ship berthing points at Waigaoqiao Harbour District, density-based spatial clustering of applications with noise (DBSCAN) algorithm is adopted. The neighborhood radius (ε) and density (MinPts) could be set according to the cluster center density and quantity of clusters. Density clustering is carried out on all berthing ships, and the clustering result is presented in figure with clusters and noises. Compared with the distribution diagram of berths and anchorages at Waigaoqiao Harbour District, a list of suspicious berthing ships is generated. By analyzing historical trajectories of ships in the list, we could make ships' real berthing records clear, and identify anomalous berthing ships at Waigaoqiao Harbour District. The study has found that the abnormal berthing ships at Waigaoqiao Harbour District were located at the Nangang Channel between Yuanyuansha Anchorage and Wusongkou Anchorage or Nangang Channel near Jiangya Nansha Anchorage. What's more, the changes of ships' position before and after berthing position small, while ships' speed before and after berthing position reduced sharply. Therefore, we could speculate that it was ship emergency failure that leads to ships' anomalous berthing. According to the ship maritime mobile service identity (MMSI), maritime safety administration (MSA) can quickly locate the shipping company related to the ship so as to strengthen the onshore ship safety management. Anomalous ships' berthing time and position help to record the failure duration and location, which can supply the important evidence for fleet management of shipping company.

Key words: maritime silk road, AIS, berthing ships, DBSCAN, anomaly detection