Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/7342
Title: 基于概率的地图实体匹配方法
Other Titles: A probabilistic theory-based matching method
Authors: Tong, X
Deng, SS
Shi, W 
Keywords: Map conflation
Probabilistic theory
Feature matching
Multi-indicators fusion
Issue Date: 2007
Publisher: 科学出版社
Source: 測繪学报 (Acta geodetica et cartographica sinica), May 2007, v. 36, no. 2, p. 210-217 How to cite?
Journal: 測繪学报 (Acta geodetica et cartographica sinica) 
Abstract: 数字地图合并是通过同名实体匹配和合并变换技术,调整相关地物实体的几何、属性等差异,实现同一地区不同来源地图数据的集成和融合。其中同名实体匹配是极为重要的第一步,也是一个存在大量不确定性的过程,匹配阈值的选取、实体非一对一的匹配关系是匹配中的关键难题,匹配效果不佳或出现错误匹配直接影响着后续合并结果的正确性。本文提出一种基于概率理论的匹配模型,该模型融合多种匹配指标,通过计算实体匹配概率大小来确定匹配实体。该方法避免了匹配指标精确阈值的选取,并且能够有效地解决匹配中非一对一的情况。
The conflation of geographic datasets is one of the key technologies in the front research area of spatial data capture and integration in Geographic Information Systems(GIS).Map conflation is a complex process of matching and merging map data.Because various reasons relate to map data discrepancies,a great amount of uncertainties exist during the process.In the first step,selecting appropriate thresholds and handling one-many or many-many matching relationships are two difficulties in feature matching,which predetermines following map merging step.This paper proposed a probabilistic method for feature matching,which fuses a variety of criteria to calculate the matching probability.The feature pair with the highest probability can be determined to be matched.This method avoids selecting thresholds and attempts to resolve one-many and many-many matching relationship.
URI: http://hdl.handle.net/10397/7342
ISSN: 1001-1595
Rights: © 2007 中国学术期刊电子杂志出版社。本内容的使用仅限于教育、科研之目的。
© 2007 China Academic Journal Electronic Publishing House. It is to be used strictly for educational and research use.
Appears in Collections:Journal/Magazine Article

Files in This Item:
File Description SizeFormat 
Tong_Probabilistic_Theory_Base.pdf1.08 MBAdobe PDFView/Open
Access
View full-text via PolyU eLinks SFX Query
Show full item record

SCOPUSTM   
Citations

37
Last Week
0
Last month
Citations as of Oct 16, 2017

Page view(s)

184
Last Week
7
Last month
Checked on Oct 22, 2017

Download(s)

62
Checked on Oct 22, 2017

Google ScholarTM

Check



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.