Abstract:
Objectives: Frequent, large-scale natural disasters inflict substantial harm on modern society, posing serious challenges to emergency decision-making, response, and assessment. During the emergency response processes, geospatial disaster data rapidly expands across physical, social, and information spaces. Despite extensive research on disaster geospatial big data, studies have yet to integrate into a unified intelligence system. As a result, fragmented disaster geospatial data fails to reflect its inherent information value. Converting complex disaster geospatial big data into emergency rescue geospatial intelligence (ER-GeoINT) is an emerging trend in modern emergency rescue. Methods: Firstly, in response to the current challenges such as the difficulty in emergency rescue data integration, the ambiguous application system of emergency intelligence, and the low efficiency of emergency rescue decision-making, we build on existing geospatial data integration processing and analysis methods. By examining the structure and characteristics of current disaster spatiotemporal big data, we propose the concept, features, architecture, sources, and quality metrics for ER-GeoINT. This clarifies the information sources and classification system of ER-GeoINT. Secondly, to address technical challenges in mining and analyzing disaster geospatial big data, we develop a system framework for ER-GeoINT generation technology that supports multiple disaster scenarios. This framework elaborately presents the current research on data perception aggregation, intelligent extraction and analysis of intelligence information, and the integrated intelligence generation technology of spatial intelligence. Results: Focusing on emergency rescue scenarios for typical natural disasters such as forest and grassland fires, earthquakes, and geological hazards, a multi-hazard supported Intelligence Collection and Analysis Service System (ICASS) has been developed. This system integrates geospatial intelligence mining technologies and methods for disaster online monitoring and early warning, multimodal emergency intelligence information extraction, rapid disaster loss assessment, and on-site situation simulation and deduction. Through these integrations, an all-source intelligence database, intelligence collection and analysis service system, and an application demonstration pattern for emergency rescue intelligence demand services have been established. Conclusions: The sources and types of global disaster emergency data are increasingly expanding, while the demands for emergency rescue intelligence are shifting towards more intelligent, spatial, granular, and personalized approaches. Current challenges and opportunities for ER-GeoINT include developing a credible and precise intelligence quality evaluation system, strengthening the "human-in-the-loop" intelligent rescue intelligence generation technology system, enhancing the fine-grained element spatial perception, grid-based intelligence collaborative management, deepening the exploration of value-added spatial intelligence through three-dimensional emergency mapping services and improving the integration with disaster emergency response plans.
下载:
