图书情报知识 ›› 2021, Vol. 38 ›› Issue (3): 61-73.doi: 10.13366/j.dik.2021.03.061

• 专题:数字人文中的跨学科融合实践 • 上一篇    下一篇

基于图情景感知计算的舆图信息元学习框架

翁子扬,王帮超,王文慧   

  • 出版日期:2021-05-10 发布日期:2021-07-03

A Meta-learning Framework of Map Information Based on Image Scenario Aware Computing

  • Online:2021-05-10 Published:2021-07-03

摘要: [目的/意义]顾及国家社会发展阶段与勘绘制图情景的计算推演,是空天遥感科学介入人文社会学研究的具体应用。本研究提出的舆图信息元学习框架模型,旨在实现档案活化与遥感GIS地理信息环境拓扑,从而完成舆图语义的泛在结构化叙述,并通过知识图谱建设,为关联研究提供定量计算的基础。[研究设计/方法] 构建舆图档案与史地典籍资源的融合知识图谱,实现与GIS环境下的地理交通位置映射。地理典著录分析辅助实现地理交通文献与空间位置的融合检索功能,提供舆图地理典籍文献与空间位置叠加的复合研究环境。[结论/发现]本文提出的信息元框架能实现单一知识目标与泛在外域知识的有效融合,使舆图资源分析研究实现从图档案-图数据-图情景的阶跃。[创新/价值]为舆图的自然描述语义和图像场景结构描述之间的语义划分提供了界限。针对不同类型的地图场景匹配合适自适应分析模型,能够有效提高知识发现的效率,以及空间人文科学领域中地理泛在计算过程的健壮性。

关键词: 舆图, 图情景, 信息元学习框架, 知识图谱, 地理信息系统, 空间计算

Abstract: [Purpose/Significance]The calculation and deduction that takes the stage of national social development and the scenario of surveying and mapping into consideration is an application of space remote sensing intervening in social science. The proposed map information meta-learning framework model is expected to realize the archival activation and environment topology of remote sensing GIS geographic information data, based upon which the ubiquitous structural description of ancient map semantics could be achieced. As a result, a quantitative calculation basis could be provided for association studies with knowledge graph construction.[Design/Methodology]An integrated knowledge graph of map archives and history books was built to realize geographic transportational location mapping. Description and analysis of geographical classics could help realize the function of integrated retrieval and provide a complex research environment for the composition of geographic classics and spatial location.
[Findings/Conclusion]The proposed meta-information framework is able to realize efficient integration of single knowledge objective with broad external knowledge. Finally, it enables the analysis and research of map resources to achieve a leap from map archives to map data, then to map scenarios.[Originality/Value]This study provides a circumscription of semantic division between natural description semantics and image scenarios structural description. Besides, it can match suitable adaptive analysis models for different types of map scenarios, which effectively improves the efficiency of knowledge discovery and the robustness of geographic ubiquitous computing process in the field of space humanities.

Key words: Map, Map scenario, Meta-learning framework of information, Knowledge graph, Geographic information system, Space calculation