改进QuickBundles算法在船舶轨迹聚类中的应用

Application of improved QuickBundles algorithm in ship trajectory clustering

摘要: 对船舶AIS数据聚类进行研究，可以挖掘出船舶航行过程中有效或潜在的信息，对于提高船舶海事交通管理和水路交通运输的智能化水平具有重要意义。传统的聚类算法在面对大量的AIS数据样本时通常表现出很低的执行效率。因而，提出一种改进QuickBundles算法，并对船舶轨迹采样方法和距离度量方式进行改进，选取长江南京航段板桥汽渡水域的船舶AIS数据作为试验依据，最终实现船舶轨迹的有效聚类。试验结果表明，与原QuickBundles算法和DBSCAN算法相比，改进QuickBundles算法在算法执行效率和算法准确性方面优于前两种算法，证明改进QuickBundles算法可有效应用于船舶轨迹聚类。

Abstract: An improved QuickBundles algorithm is introduced to improve the efficiency of massive AIS data clustering. The methods for ship trajectory sampling and distance measurement are also improved. The ship AIS data from Banqiao Ferry area of the Yangtze River is processed and effective trajectory clustering is achieved. The experiment proves the advantage of the improved QuickBundles algorithm over conventional QuickBundles algorithm and DBSCAN algorithm in efficiency and accuracy.