Integrated Data-model-knowledge Representation And Application For Natural Resource Entities

Oriented to "systematic management of mountain,water,forest,arable land,lake,grass","all-round,the whole region and the whole process of construction of ecological civilization" and other countries strategic needs.Natural resources such as mountains,water,forests,arable land,lakes,grass,sand,iceberg and other life community elements are scattered and independent,lack of organic correlation,cannot meet the integrity of geographical elements and systematic management of major problems.It is urgent to establish an integrated representation model for natural resources that explicitly describes the mutual feed-back relationships between various elements,so as to support high-level diagnostic,predictive and decision-making analysis of data-driven,model-driven and knowledge-driven cooperation,and realize the transformation from single-factor decentralized management of natural resources to all-factor system governance.Existing GIS data models focus on the description of spatial structure and temporal characteristics of the entity,but lack the overall expression of geometry and time,and pay attention to the spatio-temporal changes inside a single entity.For the description of the relationship,spatio-temporal data model tend to separate description of temporal and spatial relationships,ignore the natural resources entities internal and entity relationship between time and space,in terms of natural resources in the complex interaction relationship expression more show weakness,this means that the interconnection and interaction between natural resource entity objects and can not fully reflect the dynamic evolution process.It lacks data and information sharing mechanism,knowledge reasoning and model calculation to analyze complex spatio-temporal phenomena,difficult to meet the needs of refined and comprehensive management of natural resources,global diagnosis and decision-making.This paper establishes an integrated data-model-knowledge representation model for the needs of aboveground-surface-underground representation and management of natural resources.The specific research contents are as follows:(1)Aiming at the defects of the existing natural resource management model in the expression and display of 3D physical spatial structure relationships,a unified expression model of physical geographical elements including geometric layer,expression layer and semantic layer is constructed to achieve a unified description of three-dimensional physical data at multiple detail levels.The classification system of natural resource entities oriented to natural resource management is established.The typical spatio-temporal interactional relationships are defined.For the storage and management of massive heterogeneous natural resource data,integrated data-model-knowledge is proposed,with mechanism models,knowledge graph and intelligent algorithm to efficiently organize and manage spatio-temporal big data to meet the needs of different tasks.(2)For specific tasks in the management of natural resources efficient access to natural resource related data,reducing redundant data load,reducing the interference of the irrelevant data to the actual analysis and decision,based on natural resource management task attention,a conceptual model for multi-scale adaptive unified expression is designed to provide relevant data for efficient access to natural resources,make an accurate analysis and decision-making basis.(3)In order to break through the limitations of existing GIS data models,which are mainly descriptive analysis and lack of diagnostic and predictive intelligence services,and considering characteristics of multi-source heterogeneous data,a unified physical expression logical model of natural resources is constructed.The model codes the entities of natural resource data,designs the logical structure of integrated "data-model-knowledge",organizes and manages the entity features of natural resources and their associated relationships based on graph model to calculates and simulates complex interactions through mechanism model.A knowledge graph for natural resource management was constructed to support semantic modeling,behavior description and simulation analysis,and a query mechanism for analysis and exploration task was designed based on the knowledge graph.The optimization of query results of pattern matching was realized under semantic constraints.In order to verify the feasibility and effectiveness of the models and methods,typical natural resource entities were selected to construct natural resource entities according to the integrated data-model-knowledge representation proposed in this paper.Construct the natural resource knowledge graph.Then,the effectiveness of the multi-scale unified representation of natural resource entity geometry and semantics based on the attention of natural resource management task is verified by combining river entity.Design at the same time natural resources data model-knowledge expression system integration architecture,discuss the framework of natural resources integration,refinement and typical application of integrated management processes,explicitly expressed by natural disasters and the interaction relationship between natural resources,and human engineering activities on the natural resources and the influence of spatio-temporal relationships.It lays a foundation for efficient,accurate and transparent perception of the distribution of complex natural resources above and under the ground.