摘要
本文针对空间面群目标提出了一种几何相似度计算模型。首先,利用拓扑关系概念领域图定义了面群之间的拓扑关系相似度;然后,对不同类型的面状目标选用合适的"降维"方法处理为"线群"目标,利用方向均值定义线群之间的方向关系即面群目标的方向相似度,以及利用"环形方差"定义线群目标之间的距离关系即面群目标的距离相似度。最后,结合面群的长度和平均长度、面积和平均面积,面密度及紧致度,建立了面群目标几何相似度计算模型,以对面群目标相似度进行整体度量。该模型综合考虑了空间面群目标的几何特征和空间关系特征,并对其作了适当的权重分配。从时间邻近度和尺度邻近度角度,本文设计了2个实验,结果表明,相似度计算结果与地物特征比较一致,符合人们的直观空间认知。
As a kind of spatial relationship, spatial similarity relationship is still in its initial research stage. Re- search works of similarity assessment of spatial group objects will improve spatial relationship＇ s theory, deepen spatial cognition and raise the level of spatial data intelligent and automated handling. Focus on the spatial poly- gon groups, this paper proposed a novel computational model of geometry similarity measurement between spa- tial polygon groups. The conceptual neighborhood network of topological relationship was utilized to define to- pological relationship similarity between polygon groups. A suitable ＂dimensionality reduction＂ approach was utilized to process different polygon objects into line group. Then, the directional mean was utilized to define di- rection relationship between line groups, namely direction similarity of polygon groups. Finally, the circular vari- ance was utilized to define distance relationship between groups. Combined with the length, average length, area, line groups, namely distance similarity of polygon average area, density and compactness of polygon groups, the final step established a computational model of similarity to comprehensively measure the geometry similarity of polygon groups. The computation model considered both geometry features and spatial relationship features of polygon groups, and a proper weights distribution was taking into account at the same time. Two ex- periments were taken out to verify the model. The experimental results showed that the computed similarity is consistent with ground object features and intuitive cognition of human beings.
出处
《地球信息科学学报》
CSCD
北大核心
2013年第5期635-642,共8页
Journal of Geo-information Science
基金
国家自然科学基金项目(41201476)
兰州交通大学青年科技基金项目(2012001)
关键词
空间相似性
空间面群
几何相似度
空间关系
spatial similarity
polygon group
geometry similarity
spatial relationship
