一种三维虚拟仿真洪水预测与预警方法及系统
By acquiring topographic and rainfall data to generate flood evolution and peak prediction results, and using convolutional recurrent neural networks and conditional generative models for 3D presentation and early warning, the problem of high modeling complexity and easy error accumulation in existing technologies is solved, and rapid and accurate flood prediction and early warning are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHONGSHUI SANLI DATA TECH CO LTD
- Filing Date
- 2026-04-28
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for flood inundation evolution prediction and flood peak forecasting suffer from high modeling and computational complexity, time-consuming numerical simulation calculations, and easy error accumulation. In particular, nonlinear predictions lack reliability under extreme rainfall scenarios.
By acquiring topographic data and rainfall time series data of the target area, multi-time step prediction results of flood evolution and flood peak prediction results are generated. The results are then presented in three dimensions on a three-dimensional visualization platform, and threshold rules are used to output early warning information. A convolutional recurrent neural network (ConvLSTM) is used for spatiotemporal feature sequence prediction, and a conditional generation model is combined to predict the peak value.
It enables rapid and accurate prediction of flood evolution and peak values under extreme rainfall scenarios, and can display the evolution of disaster and risk level under a unified data and model framework, which facilitates emergency command and risk assessment.
Smart Images
Figure CN122113685B_ABST