Method Research On Construction Of Ship Activity Graph And Visual Analysis

With the deepening of global economic integration,maritime ship activities are becoming more and more frequent.Relevant regulatory authorities and various open-source channels have accumulated a large amount of ship trajectory data and maritime traffic event text data,which provides a huge space for data-driven technological innovation for the analysis of maritime ship activities.At present,ship activity analysis mainly depends on ship trajectory data,but ship trajectory data is very sparse and has the problem of low knowledge density,which can not support people to analyze complex behaviors.At the same time,without very professional knowledge,it is difficult to quickly select appropriate analysis methods,face the problem of "cold start" and low analysis efficiency.Therefore,how to use open source and multi-source ship activity data to enhance the semantic understanding of ship activity data from sparse trajectory data,realize the automatic recommendation of analysis methods,and promote the intelligent development of ship activity analysis is a major challenge.In order to solve the above problems,the main research contents of this paper are as follows:1.Aiming at the problem of sparse ship activity trajectory data and low knowledge density,this paper integrates open source and multi-source ship activity data to build a ship activity knowledge graph:(1)Preprocessing multi-source ship activity data,including geographic vector data,ship trajectory data and ship event text data;(2)The process of constructing the knowledge graph of ship activity is described.Based on the core idea of "process-event-action",the SEM(Simple Event Model)Model is improved and the ship activity ontology Model is designed.(3)Semantic information of trajectory was extracted by Stop/Move model and geographic association relationship,and relationships between ship emergencies and events were extracted by deep learning model to complete case layer filling.Experiments show that the ship activity knowledge graph constructed in this paper supports the knowledge representation of ship routine activities and emergencies,achieves the effect of semantic enhancement,and can query and trace ship activities in time and space,which has certain application value.2.Aiming at the problem of "cold start" in ship activity analysis,a ship activity visual analysis model is designed in this paper.(1)A ship activity visual analysis model involving 14 analysis scenarios is designed based on the built ship activity knowledge graph.(2)Using human-computer interaction methods such as "filter","association" and "navigation" for visual analysis,and design the flow of ship activity visual analysis;(3)The visual analysis methods of ship activity space,time sequence,time and space and attribute characteristics are summarized.Experiments show that the visual analysis model of ship activity in this paper can recommend a relatively appropriate visualization method for users,give full play to the core role of human in the visual analysis process,provide more perspectives for the analysis of ship activity,but also verify the conclusion that the change of ship activity law caused by special circumstances such as fishing ban.3.Based on the above methods and open source data,a ship activity visual analysis system is developed to trace and analyze ship activities.The main work includes:(1)Summarize the analysis requirements of five aspects of ship activity query,spatio-temporal backtracking,spatial distribution law,temporal change law and relationship characteristics,and design two functional modules: one is the query of ship activity knowledge graph,the other is the visual analysis of ship activity;(2)The realization environment of the system is described,and the system framework and main interface are designed.(3)Based on the case of ship activity knowledge atlas,ship activity was analyzed by visual analysis method in this paper,and corresponding conclusions were drawn.