【作者】 王晓莉；

【导师】 张锦；

【作者基本信息】 太原理工大学 ， 地图制图学与地理信息工程， 2007， 硕士

【摘要】 在城市地理信息系统数据库建设和空间分析应用中，经常需要用到同一地区不同比例尺的地理空间数据。但是由于这些数据往往都存在着不同程度的位置和几何形状的差异，所以我们需要进行数据整合才能更好的满足应用需求。数字合并——Conflation是实现多尺度空间数据整合的关键技术。居民地和道路网是地形图的主要表示内容，所以居民地和道路网的合并计算是地理空间数据库整合的重要方面。基于Conflation的多尺度居民地和道路网的整合研究对建立地理空间数据库合并的主要技术框架有着重要的意义，它将增强并丰富空间数据库的服务能力。本文首先利用ArcGIS的空间平差方法(仿射变换方法、边缘匹配方法以及Rubber Sheeting技术)和数字合并工具实现了某矿区两种不同比例尺地形图中同名居民地和道路网实体的几何合并和属性合并，并给出了ArcGIS的Conflation应用实例。实例表明ArcGIS的数字合并方法在一定程度上解决了不同尺度居民地和道路网的整合问题，对实际应用有一定的指导作用。通常数字环境下的居民地和道路网合并遵循自动探测同名实体和同名实体匹配的合并过程，所以匹配技术是关键。本文在分析最小二乘原理的基础上提出数字合并最小二乘匹配方法。其中详细介绍了基于X、Y坐标的最小二乘坐标匹配方法以及基于两点距离函数和P诺码理论的最小二乘点匹配方法，并依托实例对最小二乘坐标匹配方法进行了质量评定。实例表明当位于两幅不同比例尺地形图中的同名居民地群之间服从线性变换的关系时，利用坐标匹配一阶变换的方法就可基本实现两个居民地群的匹配；而当它们之间是非线性关系时，须进行高阶多项式坐标匹配或分区域匹配才能达到较好的匹配效果。最后总结论文的研究内容及结论，并提出不足之处和进一步改进和完善的方面。更多还原

【Abstract】 Geospatial datas that come from the same region, are at different scales, are needed when constructing database and doing spatial analysis In city GIS.But they usually have diverse position and figure differences so we need to integrate them.Conflation is an effective multi-scale spatial data integration strategy.The conflation of settlements and roads which are the main contents in the relief map is the pivotal problem during the process of integrating geospatial data .The integration of multi-scale settlements and roads based on conflation is very valuable for upbuilding technology framework of conflation of geospatial database and ultimately will enhance the service ability of spatial database.First, this paper integrates figure and attribute datas of identical settlements and roads in the relief map of two kinds of scales using spatial adjustment and conflation tools of ArcGIS, and the examples of conflation are given.It is indicated that the conflation of ArcGIS integrates different-scale settlements and roads and guides practice to some extent.Usually, conflation under digital circumstance follows the process which includes exploring and matching identical entity automatically, so matching is the most important.Then this article brings forward Conflation Least Squares Matching according to Least Squares Principle. Coordinates matching model based on X, Y coordinates which have been assessed by examples and point matching model based on distance function and P code theory are get.It is indicated that when the identical settlements obey linear relationship, single coordinates matching can match settlements basically but when non-linear, only high-level polynomial coordinates matching or divisional matching can match settlements well.In the end, the paper summarizes the content and conclusions of the subject and brings forward the shortage and betterment.更多还原