A Research On Construction Method Of Behavior-Event Geographic Spatio-temporal Data Model Based On MongoDB

In the era of big data,the storage and indexing requirements for geospatial data are getting higher and higher.Spatio-temporal data modeling is the basis for the temporal and spatial information system(TGIS)to reasonably and effectively express the temporal and spatial changes of spatial entities in the real world and to organize their spatio-temporal relationships.In recent years,it has been the frontier problem in the field of GIS data modeling.At present,the research of scholars and experts at home and abroad mainly focuses on the three dimensions of time,space and attribute of the geographic feature model framework of multidimensional space-time characteristics.The data organization and storage methods restrict the mining and analysis of multi-dimensional geographic data features.The event-based geographic data model has certain limitations in historical backtracking and future forecasting.It cannot describe the causal relationship between geographic entities before and after the event and the historical reasons and future development trends of geographical phenomena.Therefore,this paper first describes the multi-granular expression of complex objects composed of simple objects.In the analysis of the existing event-based model and the logical relationship between behavioral cognition and events,design behavior-event-based geographic spatiotemporal data model.At the same time,using the characteristics of MongoDB document storage,MongoDB non-relational database is designed to store the geospatial data model to realize effective management and analysis of data.This article specifically studied the following aspects:(1)The three basic geographic information granularities of semantic granularity,time granularity and spatial granularity are analyzed.On the basis of this,the description of multi-granularity organization and expression from simple object to complex object is discussed,and the relationship between measurement,orientation and spatio-temporal topology is discussed.Provides theoretical support for behavior-event based geospatial data models.(2)Analyze the attribute relationship between events and time,and analyze the hierarchical relationship between geographic entity behavior cognition and events.From the perspective of time and space cognition,explore the continuity of behavior cognition on the time axis,and design the concept of geospatial data based on behavior-event.Model and represents the logical relationship between classes and classes in the conceptual model.The model considers a behavioral cognition as a combination of multiple events,the attributes of the event and the spatial information continuously on the time axis,and designs the B-E geographic model of the "time,space,attribute,event" dimension.(3)The "non-modal" MongoDB database fragmentation clustering mechanism is studied.Based on the BE geographic data model,the geo-time-space data model based on MongoDB and the storage process based on GridFS are designed,and the data equalization algorithm is improved for the massive migration of geographic data.Combine the Hilbert curve on the original R-tree index to improve the indexing efficiency of MongoDB.From the perspective of geographic object behavior cognition and event semantics,this paper describes the causal relationship between geospatial and temporal objects by constructing a behavior-event-based spatio-temporal data model,and predicts the future of geography based on the causes and trends of change.Organize time,space,attribute and event information with object-oriented ideas,and combine MongoDB's database storage to improve data balance and indexing methods to meet the requirements of data storage,processing and mining of spatio-temporal big data.