基于DBSCAN的港口泊位自动识别算法设计

为更好地监测船舶动态和船舶在港口的作业情况,通过对K-Means算法和DBSCAN(Density-Based Spa-tial Clustering of Applications with Noise)密度聚类算法进行对比,选择DBSCAN密度聚类算法对港口泊位进行聚类,对港口泊位的位置和大小进行识别.基于船舶自动识别系统(Automatic Identification System,AIS)历史数据、船舶动态信息和船舶基础信息,采用DBSCAN密度聚类算法对全球4 079个港口的泊位进行自动识别,得出泊位的位置、方向、岸线长度和类型等信息.将聚类结果与真实泊位信息相对比,结果表明,聚类结果的误差很小,证明该算法是有效的.计算结果可用于实时跟踪船舶在港作业动态,分析泊位历史作业情况,为港航基于数据的协同优化提供参考.

Port Berth Identification with DBSCAN Clustering Algorithm

AIS data are used to find position and size information of berths in ports by means of cluster analysis.K-Means algorithms and DBSCAN(Density-Based Spatial Clustering of Applications with Noise)algorithm are compared and the latter is selected for this purpose.AIS(Automatic Identification System)data from 4079 ports worldwide,which reflect ship movement and ship basic information around the ports,are processed and berths of the ports are identified.The location,orientation,shoreline types and length of those berths are calculated and checked with the actual data of the berths,which proves the effectiveness of the algorithm.