语义尺度影响的相离区域合并与拓扑关系计算

Disjoint Region Merging and Topological Relation Computing Induced by Semantic Scale

  • 摘要: 拓扑关系随着语义尺度的变化需要重新推理或者计算。当粗略语义尺度下的区域对象由详细尺度下的有限个区域合并而成时, 区域对象间的拓扑关系可采用已有的组合推理方法得到, 然而这些方法只适用于简单对象并存在多解性。针对此问题, 提出了基于9交模型的拓扑关系多尺度计算方法, 分别针对相离区域合并和相邻区域合并定义了9交矩阵操作算子, 可利用详细语义尺度的拓扑关系直接计算出合并区域间的9交矩阵。利用9交矩阵操作算子得到的计算结果值域为复杂区域对象间所有可能的拓扑关系, 且不存在多解性, 通过消除歧义性还可扩展9交矩阵操作算子, 适用于多个相离简单区域组合的复杂区域。

     

    Abstract: Topological relations vary with the changes of semantic scales.Complex regions with diffe-rent semantic scales are merged by a finite number of simple regions and their topological relations have to be updated.The current methods make use of inferred combinatorial tables based on basic topological relations between simple regions.However, these methods are generally only applicable to simple objects and have multiple solutions.This paper presents a computation approach of modeling scale dependences of topological relations based on the 9-intersection (9I) model.In terms of the disconnected region merging and adjacent region merging, 9I-based matrix operators are defined for computing directly topological relations of the coarse semantic scale from the relations of the detailed scale.The computation results of the 9I-based matrix operators have no multiple solutions, and the computation domain can cover for all possible topological relations between complex regions.The 9I-based matrix operators can be extended to process composite regions composed of disconnected simple regions by eliminating ambiguities.

     

/

返回文章
返回