测绘学报 ›› 2021, Vol. 50 ›› Issue (6): 789-799.doi: 10.11947/j.AGCS.2021.20200191

• 地图学与地理信息 • 上一篇    下一篇

面向位置聚合的泛在地图信息分类模型

王思, 王光霞, 田江鹏   

  1. 信息工程大学地理空间信息学院, 河南 郑州 450052
  • 收稿日期:2020-05-13 修回日期:2020-11-30 发布日期:2021-06-28
  • 通讯作者: 田江鹏 E-mail:tjpeng2011@163.comc
  • 作者简介:王思(1992—),女,博士生,研究方向为地图学与地理信息系统。E-mail:plaieu_wangs@163.com
  • 基金资助:
    国家重点研发计划(2017YFB0503500);国家自然科学基金(41701457;41671407)

Classification model of ubiquitous map information facing location-based aggregation

WANG Si, WANG Guangxia, TIAN Jiangpeng   

  1. Institute of Geographical Spatial Information, Information Engineering University, Zhengzhou 450052, China
  • Received:2020-05-13 Revised:2020-11-30 Published:2021-06-28
  • Supported by:
    The National Key Research and Development Program of China (No. 2017YFB0503500);The National Natural Science Foundation of China (Nos. 41701457;41671407)

摘要: 地理信息分类是地图学的核心内容。随着泛在信息社会的来临,地理信息逐渐呈现时空泛在的新质特征,传统地理信息分类模型面临新的挑战。本文以泛在地图信息为研究对象,面向位置聚合应用需求,提出由“实例层→特征层↔维度层↔主题层”4个层次构成的信息分类模型。本文设计了一种泛在地图信息分类建模的验证方法,通过基于特征系统和信息维度的泛在地图主题特征标注,实现泛在地图信息在向量空间统一表达,并结合层次聚类算法生成泛在地图信息分类分级体系。通过气象主题分类试验对分类模型的可行性进行了验证。该分类模型本质上是一种认知规律约束下数据驱动的分类分级体系自动建模,特征层的扩展改变了传统地理信息分类模型的“实例→维度↔主题”结构,使得泛在地图信息分类在保持层级化认知结构的同时,具备细粒度语义描述能力。

关键词: 信息分类, 泛在地图信息, 位置聚合, 信息维度, 层次聚类

Abstract: Geographic information classification is the core content of cartography. With the coming of the ubiquitous information society, “spatio-temporal ubiquitous” is gradually becoming a new qualitative feature of geographic information, which has brought new challenges to traditional geographic information classification models. This paper takes ubiquitous map information as its research subject and puts forward a four-level information classification model of “instances→features↔dimensions→themes” to satisfy the need for location-based aggregation application. Then it designs a verifying method for the model, which tags ubiquitous maps’ thematic features based on the feature system and information dimensions, realizes the unified expression of ubiquitous map information in the vector space, and uses a hierarchical clustering algorithm to generate the classification and grading system of ubiquitous map information. At last, it verifies the feasibility of the model through a “meteorological theme” classification experiment.In essence, the model is to automatically build a classification and grading system, which is data-driven and constrained by cognitive patterns. And its feature level has extended the “instance→dimension↔theme” structure of traditional geographic information classification models, enabling ubiquitous map information classification in fine-grained semantic description while staying in the hierarchical cognitive structure.

Key words: information classification, ubiquitous map information, location-based aggregation, information dimension, hierarchical clustering

中图分类号: