Rescue decision via Earthquake Disaster Knowledge Graph reasoning

Yifan Jiao,Sisi You

DOI: https://doi.org/10.1007/s00530-022-01002-9

IF: 3.9

2022-10-15

Multimedia Systems

Abstract:In earthquake disaster scenes, the primary task is to formulate corresponding rescue strategies accurately and quickly according to the state and material of buildings. Therefore, researchers have researched earthquake rescue decisions based on knowledge-driven. However, the existing rescue decisions are mainly based on the experience of experts, which requires a large amount of professional knowledge. Furthermore, the scenes are often very complex. Different rescue methods and equipment are usually required for different scenes. Because of the above shortcomings, we propose a novel rescue decision algorithm via Earthquake Disaster Knowledge Graph Reasoning. On the one hand, the earthquake scene is modeled through the Visual Perception module to recognize the damaged state and material of the building in the image. On the other hand, an Earthquake Disaster Knowledge Graph is constructed, and the Graph mapping is used to model the Earthquake Disaster Knowledge Graph for learning a particular vector embedding of each entity. Finally, according to the prediction of the image and the representation of the entity, the Decision Reasoning module is proposed to calculate the similarity between the state and material of the building and each target decision for reasoning the most reasonable rescue measures for specific disaster scene. For the earthquake disaster rescue, we build the Disaster Image Dataset for recognizing of state and material of the building in the disaster scene and construct the Earthquake Disaster Knowledge Graph. Extensive experiments on the Disaster Image Dataset and Earthquake Disaster Knowledge Graph demonstrate the effectiveness of the proposed approach.

computer science, information systems, theory & methods