Qualitative Spatial Reasoning And The Construction Of Spatial Knowledge Database

At present, the spatial information is making its steps towards the industrialization. It's a hot topic in both market competition and technological innovation. The socio-economic development calls for the balance between the use of natural resources and the protection of ecological environment, so the spatial information industry is strongly needed. It promotes the research on spatial information collection, processing and usage. Remote Sensing and GPS provide technical support for the collection of spatial information, and GIS creates the theoretical and technical architecture of spatial information processing and application from the quantitative side. However, spatial information is complex. Considering the quantity, spatial information is huge. Considering the data type, it includes not only number and character, but also includes image type. Quantitative research on spatial information is not enough for the development of intelligent applications with GIS, and qualitative research is necessary.The research on qualitative spatial information mainly foucued on building qualitative spatial model and qualitative spatial reasoning(QSR). The qualitative spatial model contains study on three kinds of relationship, which are topology, direction and measure. QSR is the subfield of knowledge representation and symbolic reasoning that represents knowledge about spatial domains by finite sets of named qualitative relations. One particular aim of qualitative approaches is to build model based on human common-sense understanding of space. This makes qualitative approaches useful, for instance, in human–machine interaction. This paper builds the simplified model representing spatial objects based on circle, gives the formalized representation with GML, and foucuses on the construction of spatial knowledge database. The main work is listed as following:(1)Based on the current study on qualitative spatial relationship, this paper considers the spatial objects as circles, that's the simplified model in this paper. Every object was defined with central coordinate and radius. This paper learnt the topology, direction and measurement relationship with simplified model. It's a way translating the quantitative data into qualitative knowledge.(2)This paper learnt about the research on QSR, and talked about the topology reasoning and direction reasoning based on combinative table.(3)The traditional GIS data do not contain any qualitative knowledge, and it's not suitable for internet transmission. Before building the knowledge database we came to the spatial data format on computer. This paper gave the formalism representation mechanism with GML, which meet with the standard XML format and exchangeable on internet.(4)With the simplified model, QSR that based on combinative table and GML data defineation, this paper described building qualitative spatial knowledge database(KB) system. Talked about the architecture of KB system, the structure of database, and how to fetch knowledge. This paper analysised the knowledge capacity need for KB through experiment. An improved fusional reasoning method was proposed. And last showed the example using KB under Microsoft .Net.