多尺度地图空间居民地语义相似度计算方法

地图综合的本质是一种空间相似变换，制图者在相似原则的指导下实施概括，读图者从包含相似性的地图中形成心象地图、重构现实世界。因此，多尺度地图空间中的相似关系研究非常重要。然而，由于相似的可计算性差，且其计算的目的在于揭示更深层次的信息，地图综合中相似关系尤其是语义相似关系的研究相对较少。针对这一问题，本文以语义功能区约束下的大比例尺街区式居民地合并（1∶1750至1∶4000）为例，基于匹配距离模型计算建筑物合并中的语义相似度，得到语义相似度在关键比例尺节点的值，并对结果进行分析、评价。试验表明，语义功能区约束下的建筑物合并符合读图者的地图认知需求，本文所述方法有助于地图更好地发挥信息传输载体的作用。

Semantic similarity measurement for building polygon aggregation in multi-scale map space

Map generalization is a process of spatial similarity transformation in multi-scale map spaces.Cartographers generalize under the guidance of the similarity principle;at the same time,map readers form mental maps and reconstruct the real world from maps containing similarity.Thus,it is of great significance to study and measure the similarity relations with respect to the scale reduces in multi-scale map spaces.However,due to the poor computability of similarity and the purpose of its computation is to reveal deeper information,there are few achievements on similarity relations especially semantic relations in multi-scale map spaces.To solve this problem,semantic similarities in city block aggregation(from approximately 1∶1750 to 1∶14000)under the constraint of semantic functional units are computed,and the method for measuring the semantic similarity is the matching-distance model based on ontology and set theory.By the experiment of different city block generalization,the semantic similarity values at key scales were obtained and the results were analyzed and evaluated.The experimental results have shown that the building aggregation under the constraint of semantic functional units is in accordance with map readers’cognitive needs.The method described in this paper is helpful for map to play a better role as a carrier of information transmission.