摘要
解决实际问题需要将多方面空间信息结合进行推理,仅考虑单方面空间信息是不够的.多方面空间信息结合推理已成为定性空间推理的一个研究热点.现有拓扑与方位结合推理工作主要集中在与基于最小外包矩形或单片方位模型的结合.方位信息描述是近似的,不适于精确推理;因此分别采用主方位模型和RCC5描述方位、拓扑信息.根据定义给出基本RCC5和主方位关系间的相互依赖及异质复合表;讨论了其上约束满足问题,得到一个路径相容算法,并分析了推理复性问题.
It is inadequate considering only one aspect of spatial information in practical problems, where several aspects are usually involved together. Reasoning with multi-aspect spatial information has become the focus of qualitative spatial reasoning. Most previous works of combing topological and directional information center on the combination with MBR based direction model or single-tile directions, which are too approximate to do precise reasoning. Different from above, cardinal direction relations and RCC5 are introduced to represent directional and topological information. Investigate the mutual dependencies between basic relations of two formalisms, discuss the heterogeneous composition and give the detail composing rules. Then point out that only checking the consistency of topological and directional constraints before and after entailing by the interactive table will result the mistakes. Based on this, an improved constraint propagation algorithm is presented to enforce path consistency. And the computation complexities of checking the consistency of the hybrid constraints over various subsets of RCC5 and cardinal direction relations are analyzed at the end.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2008年第z1期279-285,共7页
Journal of Computer Research and Development
基金
国家自然科学基金项目(60496321
60373098
60573073
60603030)
国家"八六三"高技术研究发展计划基金项目(2003AA118020)
吉林省科技发展计划重大基金项目(20020303)
吉林省科技发展计划基金项目(20030523)
关键词
定性空间推理
RCC
主方位关系
异质复合
路径相容
约束满足问题
qualitative spatial reasoning
RCC
cardinal direction relations
heterogeneous composition
path-consistency
constraint satisfaction problem
