南通职业大学学报2025，Vol.39Issue(4)：78-84,7.DOI:10.3969/j.issn.1008-5327.2025.04.014
融合积水检测与气象预测的洪涝预警方法探究
A Flood Warning Method with Waterlogging Detection and Meteorological Prediction
王梓源 1马靖凯 1刘林 1王子垚 1鲍睿1
作者信息
- 1. 江苏省骆运水利工程管理处 刘老涧闸站管理所,江苏 宿迁 223800
- 折叠
摘要
Abstract
To enable the rapid risk detection of waterlogging around flood control dams and adjacent roads before extreme weather arrives,and ensure the accurate prediction of flood,a flood warning method integrating YOLO(you only look once)water detection and GRU(gated recurrent unit)weather prediction is proposed.The YOLO model trained on FloodNet dataset is deployed on a drove platform to achieve real-time waterlogging monitoring,and the sensing results are transmitted in real time via a data link to a ground warning platform.The GRU model trained on the data from China Meteorological Administration(CMA)is developed to predict hourly rainfall intensity.By combining waterlogging sensing results with topographic factors,flood hazard index(FHI)is calculated to quantify and classify flood risks.To validate the system's effectiveness,HEC-RAS 2D(hydrologic engineering center's river analysis system 2D)is used to establish simulated datasets of multi-scenario floodwater and rainfall,and a comparative analysis is conducted between the FH I and the simulated ones.The results show that the warning system achieves 98.2%FH I prediction accuracy,demonstrating high reliability and practical value.关键词
道路积水感知/降雨量预测/YOLO/门控循环单元/洪涝预警指数Key words
waterlogging perception/rainfall prediction/YOLO/GRU/flood hazard index分类
信息技术与安全科学引用本文复制引用
王梓源,马靖凯,刘林,王子垚,鲍睿..融合积水检测与气象预测的洪涝预警方法探究[J].南通职业大学学报,2025,39(4):78-84,7.
