基于图结构的暴雨事件组织方法研究
|
摘要
在高时空分辨率的降雨数据支持下,分析和追踪暴雨事件的精细结构和动态行为成为可能,但存在数据量大、数据组织结构复杂的特点,对暴雨事件的时空组织与存储提出了巨大挑战.本文基于暴雨事件的演化过程,设计了一种"顶点-边界"的图结构表达模型,并构建基于Neo4j的图数据库,实现暴雨事件的组织与存储.图数据结构中的顶点描述暴雨事件对象,记录暴雨事件空间位置、降雨数据等状态信息;图有向边对暴雨顶点进行关联,记录暴雨状态之间的关系、变化速度、方向等动态信息.最后,综合时间序列模型和空间拓扑关系,设计基于图的暴雨事件组织模型,基于Neo4j图存储方式R-Tree空间索引存储暴雨矢量数据集,并构建暴雨事件库.基于Neo4j和Oracle Spatial的暴雨事件库的对比分析结果表明,Neo4j事件库在数据入库和时空查询性能方面具有优势.
Supported by the rainfall remote sensing products with high spatiotemporal resolution, it is possible to analyze and track the fine structure and dynamic behavior of rainstorm events. It is a great challenge to organize and store the rainstorm events because of the characteristics of large amount of data and complex data organization structure. Based on the evolution process of the rainstorm event, this paper designs a representation model of graph structure based on vertex-boundary, and builds a graph database based on Neo4 j to achieve the organization and storage of the rainstorm events. The rainstorm events object is described by the vertex in the graph structure, and the state information of rainstorm events such as spatial location and rainfall data are recorded as vertex's attributes. Using directed edge to correlate rainstorm vertexes, the relationship, speed change and direction of rainstorm states are recorded as directed edge's attributes. Finally, the paper designs the graph-based rainstorm events organization model based on the time series model and spatial topological relationship. The rainstorm vector dataset is stored based on R-Tree spatial index that is the Neo4 j graph storage method, and the rainstorm events database is constructed. The comparative analysis results of the rainstorm events library based on Neo4 j and Oracle spatial show that the Neo4 j rainstorm events library has advantages in data storage and spatiotemporal query performance.
Supported by the rainfall remote sensing products with high spatiotemporal resolution, it is possible to analyze and track the fine structure and dynamic behavior of rainstorm events. It is a great challenge to organize and store the rainstorm events because of the characteristics of large amount of data and complex data organization structure. Based on the evolution process of the rainstorm event, this paper designs a representation model of graph structure based on vertex-boundary, and builds a graph database based on Neo4 j to achieve the organization and storage of the rainstorm events. The rainstorm events object is described by the vertex in the graph structure, and the state information of rainstorm events such as spatial location and rainfall data are recorded as vertex's attributes. Using directed edge to correlate rainstorm vertexes, the relationship, speed change and direction of rainstorm states are recorded as directed edge's attributes. Finally, the paper designs the graph-based rainstorm events organization model based on the time series model and spatial topological relationship. The rainstorm vector dataset is stored based on R-Tree spatial index that is the Neo4 j graph storage method, and the rainstorm events database is constructed. The comparative analysis results of the rainstorm events library based on Neo4 j and Oracle spatial show that the Neo4 j rainstorm events library has advantages in data storage and spatiotemporal query performance.
引文
[1] Tu E, Kasabov N, Othman M, et al. NeuCube(ST)for spatiotemporal data predictive modelling with a case study on ecological data[C]//Neural Networks(IJCNN), Piscataway, NJ:IEEE, 2014:638-645.
[2] Hagerstrand T. What about people in regional science[J]. Papers in Region Science, 1970, 24(1):6-21.
[3] Langran G E. Time in geographic information systems[M]. Taylor and Francis, 1993.
[4] Peuquet D J, Duan N. An event-based spatiotemporal data model(ESTDM)for temporal analysis of geographical data[J]. International Journal of Geographical Information Systems, 1995, 9(1):7-24.
[5]陈志泊,陆守一.TGIS中的时空数据模型的研究进展[J].河北林果研究,2003, 18(4):395-400.Chen Z B, Lu S Y. Research advances of spatiotemporal data model in temporal geographical information system[J]. Hebei Journal of Forestry and Orchard Research, 2003, 18(4):395-400.
[6]薛存金.海洋GIS时空过程数据模型研究[D].北京:中国科学院地理科学与资源研究所,2008.Xue C J. Research of marine GIS spatiotemporal process data model[D]. Beijing:Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, 2008.
[7] Xue C J, Liao X H. Novel algorithm for mining ENSO-oriented marine spatial association patterns from rasterformatted datasets[J]. International Journal of Geo-Information, 2017(6):1-15.
[8] Xue C J, Dong Q, Xie J. Marine spatio-temporal process semantics and its applications—Taking the El Nino Southern Oscilation process and Chinese rainfall anomaly as an example[J]. Acta Oceanologica Sinica, 2012,31(2):16-24.
[9] Webber J. A programmatic introduction to Neo4j[J]. Addison Wesley Pub Co Inc, 2012:217-218.
[10]陆晓华,张宇,钱进.基于图数据库的电影知识图谱应用研究[J].现代计算机,2016(7):76-83.Lu X H, Zhang Y, Qian J. Implementation of movie knowledge graph based on graph database[J]. Modern Computer, 2016(7):76-83.
[11] Mahesh L. Neo4j graph data modeling[M]. Birmingham:Packt Publishing, 2015.
[12] Pultar E, Cova T J, Yuan M, et al. EDGIS:A dynamic GIS based on space time points[J]. International Journal of Geographical Information Science, 2010, 24(3):329-346.
[13]刘岳峰,康葳.一种基于对象快照模型的时空查询原子模型[J].北京大学学报(自然科学版),2015, 51(4):755-762.Liu Y F, Kang W. An atomic model of spatiotemporal query based on object-oriented snapshot model[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 2015, 51(4):755-762.
[2] Hagerstrand T. What about people in regional science[J]. Papers in Region Science, 1970, 24(1):6-21.
[3] Langran G E. Time in geographic information systems[M]. Taylor and Francis, 1993.
[4] Peuquet D J, Duan N. An event-based spatiotemporal data model(ESTDM)for temporal analysis of geographical data[J]. International Journal of Geographical Information Systems, 1995, 9(1):7-24.
[5]陈志泊,陆守一.TGIS中的时空数据模型的研究进展[J].河北林果研究,2003, 18(4):395-400.Chen Z B, Lu S Y. Research advances of spatiotemporal data model in temporal geographical information system[J]. Hebei Journal of Forestry and Orchard Research, 2003, 18(4):395-400.
[6]薛存金.海洋GIS时空过程数据模型研究[D].北京:中国科学院地理科学与资源研究所,2008.Xue C J. Research of marine GIS spatiotemporal process data model[D]. Beijing:Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, 2008.
[7] Xue C J, Liao X H. Novel algorithm for mining ENSO-oriented marine spatial association patterns from rasterformatted datasets[J]. International Journal of Geo-Information, 2017(6):1-15.
[8] Xue C J, Dong Q, Xie J. Marine spatio-temporal process semantics and its applications—Taking the El Nino Southern Oscilation process and Chinese rainfall anomaly as an example[J]. Acta Oceanologica Sinica, 2012,31(2):16-24.
[9] Webber J. A programmatic introduction to Neo4j[J]. Addison Wesley Pub Co Inc, 2012:217-218.
[10]陆晓华,张宇,钱进.基于图数据库的电影知识图谱应用研究[J].现代计算机,2016(7):76-83.Lu X H, Zhang Y, Qian J. Implementation of movie knowledge graph based on graph database[J]. Modern Computer, 2016(7):76-83.
[11] Mahesh L. Neo4j graph data modeling[M]. Birmingham:Packt Publishing, 2015.
[12] Pultar E, Cova T J, Yuan M, et al. EDGIS:A dynamic GIS based on space time points[J]. International Journal of Geographical Information Science, 2010, 24(3):329-346.
[13]刘岳峰,康葳.一种基于对象快照模型的时空查询原子模型[J].北京大学学报(自然科学版),2015, 51(4):755-762.Liu Y F, Kang W. An atomic model of spatiotemporal query based on object-oriented snapshot model[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 2015, 51(4):755-762.