基于时空相似性的跨域目标关联分析方法

多业务域目标聚合是情报分析的核心问题之一，为解决跨域目标关联分析工程化实现，提出了一种时空相似计算的分析方法。首先，对接收到的目标时空数据进行处理，形成元数据；其次，按照预定义的数据模型，从元数据中抽取目标信息；再次，通过时空相似度计算，对目标间的时空信息进行分析，得到目标间的时空关系；最后，利用先验知识对目标时空关系进行约束，实现从时空关系到语义关系的转换。数值仿真验证了该方法能够简单有效地实现跨域目标关联分析。

A cross-domain object correlation analysis method based on spatial and temporal similarity

Multi-service domain target aggregation is one of the key problems of intelligence analysis.In order to solve the engineering realization of cross-domain target correlation analysis,a spatio-temporal similarity calculation method is proposed.First,the target space-time data received is processed to form metadata.Secondly,the target information is extracted from the metadata according to the predefined target data model.Thirdly,by calculating the spatial and temporal similarity,the spatial and temporal information between objects is analyzed,and the spatial and temporal relation between objects is obtained.Finally,by using the prior knowledge,the target space-time relation is constrained to realize the transformation from time and space to semantic relations.Numerical simulations show that the proposed method can implement cross-domain object correlation analysis easily and effectively.