下载中心
优秀审稿专家
优秀论文
相关链接
摘要
时空数据库是研究如何存储历史和当前的时空数据,从而跟踪分析某一区域的变化,最终实现时空模型化和模拟地学过程.采用历史关系数据库模式的时空数据库可以充分利用传统时态数据库时态查询功能和GIS空间分析处理功能这些方面成熟的研究成果和现有系统,降低时空数据库建立的费用和开销.分析了采用扩展关系型时空数据库的原因,介绍了历史关系数据库模式、时态关系代数和查询语言,研究了采用历史关系数据库模式在全关系化空间数据库中组织时态信息的方法,从而实现时空数据库.以麦地轮作为例说明了采用历史关系数据库模式建立时空数据库的可行性和有效性.
Spatial-temporal database researcheson how to store historicaland currentspatial-temporaldata so as to track and analyze changes of some region, which finally mi plements spatial-temporal modeling and geographical smi ulation process. Atpresent there aremainly two solutions as for spatio-temporal database. The one is adding tmi e dmi ension on the basis of spatial database. The other is extending spatial dmi ension on the basis of temporal database.The research on spatialdatabase has alreadybeenmature at the present tmie. Sowe should fullymake use ofexisting spatial data model and spatial processing and analyzing functions of GIS, which will reduce enormously workload.Consequently this paper chooses to add tmi e dmi ension on the basis of spatial database. Extended relational spatio-temporal database extends tmie dmiension by considering tmi e as attribute ofspatialgeometry and thematic feature, which has its particular advantages. For one thing, relational database is supported by strict relational algebra and hasmany available large commercial relational databases. In addition, many information systems are originally established on the basis of relational database. In the second place many researches on temporal database itselfmostly adopt relational mode,l which extend temporal semantics for relational algebra and develop temporal structured query language. Thirdly many researches on spatial database in GIS domain also employ relational data mode,l and research on spatial query language is also based on relationalmode.l All in al,l spatial-temporal database adoptingmodel of historical relational database can make the best of research results and existing systems such as temporal query function of temporal relational database and spatial process and analysis function ofGIS, reducing the costof establishing spatial-temporal database.This paper at first introduces historical relational database mode,l temporal relational algebra and temporal query language,and then conducts researcheson historical relationaldatabase to organize temporal information in spatialdatabase.Ami ing at completely relational spatialdatabase,this papermakesuse of temporally upward compatible function of ATSQL2 and adopts themode of transplanting non-temporal database to temporal database to extend temporal information for spatial data. In the lightoforganizing principle ofspatio-temporaldata, currentdata and historicaldata are stored separately. As a result, spatio-temporaldatabase includes currentspatialdatabase and historical spatialdatabase. Currentdatabase stores the latestbase state version ofmap layers, yethistoricaldatabase storeshistoricalbase state version ofmap layers. Ifspatialdata and attribute data change, data ofnew versionwillbe added to currentspatialdatabase.At the same tmie originaldatawill become data of old version and be shifted to historical spatial database. We can adopt indexing pointer to establish the relationship between current spatial database and historical spatial database. The paper eventually gives an example of rotatingwheat fields to examine the feasibility and validity of adoptingmodel of historical relational database to establish spatial-temporal database. Rotatingwheat fields has obvious temporal trait. Generally speaking, it is not likely thatwheat fields change in aspectof spatial geometry. Most changes ofwheat fields are changes of attributes such as plant type and owner. When plant type ofawheat field changes, it is considered that the old geographic entity disappears, and a new one comes into being. We need to add a new record in the current database and transfer the original record to the historical database. In themean tmi ewe adoptPreVersion and PostVersion to record successive relationship between them. The result shows that spatial-temporal database based on historical relationalmodel is feasible and valid.However, spatio-temporal database based on historical relational database still has the shortcomings. In addition, it can notdealwith transaction tmi e yet, atwhich database objects are created, updated and deleted. Consequently in future workwe should take disposal ability of transaction tmi e to spatio-temporal database into account.