Unified Representation Method for Multi-Source Information in Situation Awareness Knowledge Graph

Knowledge graphs are instrumental for storing structured data and find wide applications in various fields like information retrieval, recommendation systems, and more. In military contexts, they serve as valuable tools for organizing structured data pertinent to the battlefield. However, conventional knowledge graphs encounter difficulties in achieving cohesive representation when dealing with the diverse and heterogeneous data streams originating from multiple sources. This paper aims to address this challenge by constructing a knowledge graph specifically tailored to military equipment, leveraging structured information available on the internet. Moreover, in order to effectively handle the diverse and heterogeneous nature of battlefield information, the paper proposes a novel knowledge graph-based model designed for the unified representation of multi-source data. This model operates by independently transforming image information and knowledge graph node information into feature vectors. By aligning these vectors, which represent image features and entities within the knowledge graph, a unified representation is attained. Consequently, this approach facilitates the creation of a Situation Awareness knowledge graph, establishing cohesive links between image data and information within the knowledge graph.