基于空间相似关系的居民地群组目标地图综合质量评价  

A Spatial Similarity Based Quality Evaluation Approach for Buildings Group Generalization

作　　者：刘涛[1,2] 段晓旗 闫浩文[1,2] 王中辉[1,2] 张黎明 LIU Tao DUAN Xiao - qi YAN Hao-wen WANG Zhong - hui ZHANG Li- ming(Faculty of Geomatics ,Lanzhou J iaotong University ,Lanzhou 73007 Gansu Provincial Engineering Laboratory for National Geographic State Monitoring ,Lanzhou 730070, Chin)

机构地区：[1]兰州交通大学测绘与地理信息学院,甘肃兰州730070 [2]甘肃省地理国情监测工程实验室,甘肃兰州730070

出　　处：《地理与地理信息科学》2017年第2期52-56,68,共6页Geography and Geo-Information Science

基　　金：国家重点研发计划资助项目(2016YFC0803106);国家自然科学基金项目(41201476;41561090);地理空间信息工程国家测绘局重点实验室开放研究基金项目(201409)

摘　　要：数字环境下的计算机自动制图综合质量评价,其本质就是计算综合前原图和综合后结果图之间的相似度,因此该文提出了基于空间相似关系对居民地群组地图要素的制图综合质量进行评价的方法。首先,根据多尺度空间相似关系的定义,提取出居民地群组多边形的特征因子并计算各因子综合前后的相似度;其次,在合理分配权重的基础上,建立居民地群组多边形总体相似度计算模型,并根据相似度计算结果对居民地群组目标要素的制图综合质量进行评价;最后,利用实验验证了该文所提评价方法的有效性。该研究为客观评价地图综合质量提供了一个新的思路。Building is a vitally important element on map. Many researches have been done on buildings generalization. However, most researches focus on algorithm design and implementation for map generalization, lacking the quality evaluation of map gen- eralization results. The quality evaluation has long been a critical point for automatic map generalization. In computer-aided auto- matic map generalization process, the essence of map generalization＇s quality evaluation is to evaluate the similarity between maps before and after generalization. Thus a spatial similarity based map generalization＇s quality evaluation approach for build- ings group is put forward in this paper. Based on the definition of multi-scale spatial similarity,a set of parameters are selected to describe buildings group, including spatial relation（mainly topological relation, direction relation and distance relation） param- eters and geometric features（mainly area, perimeter,compactness, area density and outline polygon）. The similarity assessment model of each parameter （feature） is derived accordingly. After proper weight assignment, a comprehensive spatial similarity as- sessment model is established finally. The model＇ s validity is verified by a set of experiments. This paper＇ s research affords a novel perspective for evaluating map generalization＇s quality objectively.

关 键 词：空间相似关系 地图综合 质量评价 居民地群 

分 类 号：P208[天文地球—地图制图学与地理信息工程]