人民黄河2025，Vol.47Issue(4)：44-50,7.DOI:10.3969/j.issn.1000-1379.2025.04.008
基于遥感影像的溃坝风险人口估算模型研究
Study on Population Estimation Model of Dam Break Risk Based on Remote Sensing Images
摘要
Abstract
A risk population estimation model based on Sentinel-2A remote sensing images was proposed to address the lack of downstream population distribution data in the risk assessment of life loss for small reservoirs.Taking the DYJ reservoir as an example,supervised classi-fication was used to quickly extract the building information of rural residents from remote sensing images,and combined with the evolution results of dam break flood,a risk population estimation model was built considering the flooded state of the buildings.Four supervised classi-fication methods,including Maximum Likelihood Classification(MLC),Support Vector Machine(SVM),Mahalanobis Distance Method(MDM),and Minimum Distance Classification(MDC),were compared.The results show that the MDC has an overall classification accura-cy of only 65.7%,while the other three methods achieve over 90%accuracy.Among them,the MLC provides the highest recognition accura-cy for remote sensing images,with an overall classification accuracy of 97.55%and a Kappa coefficient of 0.93,in extracting residential building elements in the study area.With the construction information extracted by MLC and the dam break risk population estimation model,the risk population downstream of the DYJ reservoir is estimated to be 1 470 people.This model can rapidly estimate the risk population downstream of small reservoirs,addressing the shortcomings in accuracy and efficiency of population density methods and cumulative estima-tion methods for residential areas.关键词
小型水库/遥感影像/风险人口/溃坝/洪水演进Key words
small reservoir/remote sensing image/risk population/dam break/flood routing分类
水利科学引用本文复制引用
余再康,程井,余雷,向思瑶,李宗樾..基于遥感影像的溃坝风险人口估算模型研究[J].人民黄河,2025,47(4):44-50,7.基金项目
水利部重大科技项目(SKS-2022153) （SKS-2022153）
国家重点研发计划项目(2022YFC3005501) （2022YFC3005501）
贵州省水利科技经费项目(KT202217,KT202218) （KT202217,KT202218）
