Flood affected population quantification assessment and attribution method, system, device, and medium

CN122286136APending Publication Date: 2026-06-26WUHAN UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
WUHAN UNIV OF TECH
Filing Date
2026-03-12
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Traditional flood disaster assessment methods are ill-suited to the increasingly sophisticated needs of current disaster prevention and mitigation efforts. They neglect the complete spatiotemporal boundaries of flood events and the spatial connectivity of hydrological events, resulting in assessment results that fail to accurately reflect the disaster-causing mechanisms.

Method used

Using multi-level watershed datasets and historical flood event data, a reverse recursive search algorithm is used to establish the spatial association between geospatial seed points and the multi-level watershed topology network. A binary physical domain mask is constructed, and the Transformer model is used for deep feature reconstruction and contribution score calculation.

Benefits of technology

It achieves accurate quantitative assessment of the flood-affected population, can identify the specific spatiotemporal boundaries and hydrological topological relationships of flood events, constructs an adaptive irregular pixel set model, and captures the disaster-causing characteristics of the entire process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122286136A_ABST
    Figure CN122286136A_ABST
Patent Text Reader

Abstract

This invention provides a method, system, device, and medium for quantitative assessment and attribution of flood-affected populations. The method includes: acquiring multi-level watershed datasets and multi-source heterogeneous historical flood event datasets; solving discrete pixels of flood-inundated area images into geospatial seed points, and using a reverse recursive search algorithm to establish spatial associations between geospatial seed points and multi-level watershed topology networks to obtain a binarized physical domain mask; performing deep feature reconstruction and spatial scale alignment on the flood event start-end time window and the binarized physical domain mask to construct a standardized multi-source disaster-causing factor feature matrix; mapping the standardized multi-source disaster-causing feature matrix to a high-dimensional embedding space to obtain a multi-source disaster-causing sequence; inputting the multi-source disaster-causing sequence into a pre-trained Transformer model to output disaster-causing factors and their contribution scores. This invention enables quantitative assessment and attribution of flood-affected populations, providing evidence-based decision-making support for post-disaster recovery planning.
Need to check novelty before this filing date? Find Prior Art