土地利用规划指标时空知识图谱的构建与可视化方法研究

On the Construction and Visualization Method for Spatial-Temporal Knowledge Graph of Land Use Planning Indicators

为解决现有土地利用规划指标库数据不全、关联性不强且规划要素感测慢等问题,本研究基于眉山市现有数据和知识图谱技术构建了土地利用规划指标时空知识图谱,使用数据库与网络爬虫技术定期获取眉山市土地利用规划指标众源信息,并结合BERT模型识别众源信息中指标的关系,补全、更新土地利用规划指标数据库,进而构建图谱模型,将实体及关系存储于Neo4j图数据库,形成眉山市土地利用规划指标时空知识图谱。在此基础上,设计一个前后端交互系统,实现了时空知识图谱及地理数据的可视化展示。系统的应用表明,构建的时空知识图谱能够实时补全数据库且能展示指标间的语义关系,可为指标及其相关数据的存储、智能化搜索及辅助土地利用规划提供技术支持。

In order to solve the problems of incomplete data,weak correlation and slow sensing of planning elements in the existing index database of land use planning,this paper constructed the spatialtemporal knowledge graph of land use planning index based on existing data and knowledge graph technology of Meishan city.And with database and web crawler technology,the paper made the multisource information of the land use planning index of Meishan city available on a regular basis.In addition,BERT model was used to identify the relationship of indicators in the multi-source information of Meishan city,as well as to complete and update the database of land use planning indicators,which allowed the knowledge graph model to be constructed.By doing so,entities and relationships could be stored in the Neo4j graph database to form the spatial-temporal knowledge graph of land use planning indicators in Meishan city.On this basis,an interactive system was designed to realize the visualization of spatial-temporal knowledge graph and geographical data.And the application of the system showed that the constructed spatial-temporal knowledge graph could complete the database in real time and showed the semantic relationship between indicators,which could provide technical support for the storage of indicators and their related data,intelligent search and auxiliary land use planning.