A typhoon disaster loss dynamic assessment method and system based on multi-source data
By combining tower monitoring data and meteorological data to build a dynamic prediction model, the problems of real-time updates and accuracy in typhoon disaster loss prediction have been solved, enabling timely and accurate assessment of typhoon disasters and improving the scientific nature of emergency management and the efficiency of resource allocation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 应急管理部大数据中心
- Filing Date
- 2026-04-24
- Publication Date
- 2026-07-17
AI Technical Summary
Existing methods for predicting typhoon disaster losses are difficult to update in real time during the disaster process, have limited prediction accuracy, and are difficult to perceive the actual impact of the disaster, resulting in delayed emergency decision-making and inaccurate resource allocation.
By acquiring real-time tower monitoring data (offline rate, power outage rate, and flooding rate), along with dynamic meteorological data and regional static data, input features are constructed. A dynamic prediction model is built using causal hollow convolutional networks and graph attention networks. The model outputs the cumulative disaster losses after the typhoon event ends and generates an interpretable emergency briefing by combining it with a large language model.
It enables real-time and accurate assessment of typhoon disaster losses, improves the accuracy of predicting casualties and economic losses, assists in the scientific and forward-looking nature of emergency management, and reduces decision-making delays and resource waste.
Smart Images
Figure CN122414554A_ABST