从时间快照视角提出了一个群体识别算法框架.首先,讨论了船舶群体的概念,并结合群体动力学和视觉感知相关理论,定量化了基本群体模式(如直线、曲线和流)的识别规则.接着,利用苏伊士运河地区的船舶AIS(automatic identification system)数据进行了案例研究.通过与 DB-SCAN(Density-Based Spatial Clustering of Applications with Noise)和局部方向中心性算法(Clustering by measuring lo-cal direction centrality,CDC)进行对比,本文方法在识别细粒度船舶群方面表现更优.
A Methodology of Vessel Group Motion Pattern Recognition in Snapshot Perspective
In this study,we propose a framework for group recognition algorithms from a time snapshot perspective.First,we discuss the concept of vessel groups and quantify the recognition rules for basic group patterns(e.g.,collin-ear,curvilinear and flow)by combining perception theory and group dynamics theory.Next,we conducted a case study us-ing automatic identification system(AIS)data from the Suez Canal area.Compared with DBSCAN(density-based spatial clustering of applications with noise)and CDC(clustering by measuring local direction centrality),the method proposed in this study performs better in identifying fine-grained ship clus-ters.
NETL
NSTL
万方数据
