空间变化是空间信息处理中的热点问题,标识变化是空间变化的关键问题之一.但是现有工作侧重表达,知识推理能力较弱.为此,研究了基于标识变化的时空推理问题.给出了基于多段图和集合论的标识变化形式化语义,在此基础上定义了二元关系模型.针对基本关系约束满足问题,提出了构造性求解算法,并证明了代数封闭算法可解;针对全关系约束满足问题,提出了一致场景算法.理论分析和实验应用均表明该方法能有效解决信息不完整、不准确情况下的标识变化表示与推理问题,对此类问题其它同类模型尚不能很好解决.
The spatial change is a hot-spot in spatial information processing. The identity-based change is one of the key problems in spatial change. The previous works focused on the representation, and can be improved in the knowledge based reasoning. So we study the identity change based spatio-temporal reasoning. Formal semantics of the identity-based change is given based on multi-stage graph and set theory. A binary relation model is defined. With regard to the basic relation, a constructive algorithm for constraints satisfaction problem is proposed and the algebra- closure algorithm is proved sufficiently. As for the full relation, a consistent scene algorithm is provided. Theoretical analysis and experiment both show this method can solve the representation and reasoning of the identity change with incomplete or incorrect information, while the previous works couldn't deal with them.