面状空间要素相似性度量方法研究  

Research on algorithms of similarity measurement for areal spatial data

作　　者：徐广翔[1] 陈杰[1] 马素媛 

机构地区：[1]山西大同大学矿业工程系,山西大同037003 [2]山西省雁北煤炭工业学校采测科,山西大同037005

出　　处：《测绘科学》2013年第3期31-33,共3页Science of Surveying and Mapping

摘　　要：不同尺度、来源的地图上同一要素通常具有一定的相似度。地图空间要素相似度在GIS领域具有广泛的应用。论文在总结前人相关成果的基础上分别从位置(距离)、形状、大小三个方面给出了面状空间要素相似性度量模型:以分形维数和面积/周长(紧凑度)作为相似特征的形状相似度;以中值距离作为相似特征的距离相似度;以面积或周长作为相似特征的大小相似度。最后,以多尺度面状空间要素为实验数据,通过比较分析验证了本文提出的相似性度量模型可行性。实验结果表明:以中值距离、分形维数作为相似度指标的度量模型综合考虑了面状要素局部结构和整体分布,在面状空间要素相似性度量方面具有很好的稳定性。The same feature from different sources or scales is usually similar on the map. The similarity measurement of map fea- tures is conducive to the clustering, query, match and update of the map in the GIS. In the paper, a reclassification of similarity meas- urement was made from the consideration of distance, shape, length based on per-researchers relative results. The corresponding simi- larity measurements were further developed, including： the fraetal dimension and area/perimeter （compactness） as the similarity of shape; the differences in median distance as the similarity of position（distance） ; the area deformation and perimeter difference as the similarity of size. And then, some practical area data in multi-scales were provided to illustrate the proposed methods in the paper. Fi- nally, the experimental resuhs showed that the model of distance similarity based on median distance and shape similarity based on fractal dimension taking into account the whole distribution and local structure for feature is stability. Key words： similarity; fraetal dimension; median distance

关 键 词：相似度 分形维数 中值距离 

分 类 号：P282[天文地球—地图制图学与地理信息工程]