地理邻域关系认知下基于粒子群优化的居民地自适应匹配方法

项目名称： 地理邻域关系认知下基于粒子群优化的居民地自适应匹配方法

项目编号： No.41301426

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 天文学、地球科学

项目作者： 叶亚琴

作者单位： 中国地质大学（武汉）

项目金额： 25万元

中文摘要： 多尺度居民地匹配是基础地理数据分析与处理的基础问题和难点问题。项目针对现有方法在尺度差异较大情况下的数据相似性表达、多相似性指标融合方法、M:N的复杂匹配情况处理等方面存在的不足，提出了相应的解决方案。项目在居民地与道路网、水系等邻域地物的地理空间关系认知下，兼顾居民地匹配的全局和局部层次性，提取符合人类认知模式的区域划分法则、建立加强地理邻域关系的特色匹配指标体系；考虑具体数据特征对于整体相似性计算的影响，提出基于粒子群优化算法的多指标自适应融合模型，解决整体相似性的计算问题；同时结合双向匹配机制和群组相似性表达解决M:N的复杂匹配。项目最终建立一套地理空间关系认知下基于粒子群优化的居民地自适应匹配方法，以提升多尺度居民地匹配的稳定性和正确性。

中文关键词： 地理认知；本体匹配；空间相似性认知；粒子群优化；

英文摘要： Multi-scale residents' matching is a basic and difficult problem in geographic data analysising and processing. Because existing match methods have several problems. For example, they are weak on dealing with those date whose scale is very different. And their fusion method of similarity indexes is not perfect. What's more, they also cannot do well with M:N complicated match processing. This project addresses these issues and wants to pay some progress on the matching. The main idea of the project are as follows. The first step, the project analysises geographical relationship of residents and their neighborhood features，such as road network, urban water system. So as to extract zoning rules in compliance with human cognitive models and build matching index system by strengthening the relationship between the characteristics of geospatial neighborhood. This step takes into account the residents match constraints from both the global and local levels. The second step, we consider the specific characteristics of the data for the overall similarity calculation, and proposed adaptive fusion model of multi-index based on particle swarm algorithm to solve the overall similarity calculation. The third step, we combines two-way matching mechanism and spatial asessment similities to resolve M:N the complexity of matching

英文关键词： Geographical Cognition；Ontology Matching；spatial similarity cognitive；particle swarm optimization；