面向按需服务的变化检测领域知识表达模型

Knowledge representation model for on-demand change detection

为了实现变化检测领域知识的有效建模与表达,继而支持变化检测处理服务链的按需灵活构建,该文在分析变化检测领域特征的基础上,借鉴超图模型对多元关系表达能力强的特点,提出了基于分层超图的变化检测领域知识表达模型。首先,将该模型定义以服务为结点,以服务关系为边的网络模型。然后,对结点集合和边集合进行了详细阐述,将结点集合分为数据服务结点和3种粒度的处理服务结点,将边集合分为无向简单边、有向简单边、无向超边和有向超边4种类型。最后,探讨了基于该模型的更新演化与按需服务组合方法,并以实际变化检测为例进行了案例分析。实验证明该模型具备很好的可扩展性,能够有效地支持变化检测按需服务生成。

In order to realize change detection domain knowledge modeling and representation,and support on-demand service chain generation.Based on change detection domain characteristics analysis,this paper proposed a layered hypergraph-based knowledge representation model(LHKRM)by extending the traditional hypergraph model.Firstly,we defined web services as nodes and service relations as edges in LHKRM.Then,we elaborated the nodes and edges.The nodes were divided into data services and three granularities of processing services.The edges were divided into four types of undirected simple edge,directed simple edge,undirected hyper edge and directed hyper edge.Finally,we discussed the approaches of model updating and on-demand service composition.A case study of land cover change detection was used to evaluate the proposed model.Experiment results showed that the LHKRM had the characteristic of extensibility,and could be effectively used in the on-demand service generation of change detection.