华中科技大学学报（自然科学版）2026，Vol.54Issue(7)：48-54,7.DOI:10.13245/j.hust.260055
数字孪生流域多尺度数据底板构建技术研究
Research on the technique for constructing multi-scale data baseplates of digital twin watershed
摘要
Abstract
To meet the needs of decision-makers in pre-simulations who require a focused analysis of the feasibility of scheduling plans across different scale scenarios,a digital twin watershed multi-scale data base construction technology based on orthorectified texture mapping,an improved semi-global matching algorithm,and an attention-connected cost stereo matching network was proposed.By combining open-source terrain data with low-cost satellite remote sensing data,the construction of data bases for different scales,including large-scale watersheds,medium-scale river segments,and small-scale engineering projects,was achieved.Taking the Yangtze River basin as an example,data bases with resolutions of 500 m for the entire basin,5 m for the section from the Three Gorges to Hankou,and 1 m for the Three Gorges and Gezhouba hydro projects were constructed.This meets the needs of decision-makers in pre-simulations who focus on analyzing the feasibility of scheduling plans from different perspectives,such as watersheds,important river segments,and hydraulic engineering projects.At the same time,it reduces the computational resources required for detailed scene rendering,providing support for intelligent simulation and precise decision-making in watershed flood control operations.The relevant technical achievements presented in this paper can serve as a reference for the construction of data bases in other watersheds.关键词
长江流域/数字孪生流域/数据底板/多尺度/半全局匹配/注意力连接代价立体匹配网络Key words
Yangtze River basin/digital twin watershed/data baseplate/multi-scale/semi global matching/attention concatenation volume stereo matching network分类
建筑与水利引用本文复制引用
周丽伟,张东映,韩轶龙,李晓东..数字孪生流域多尺度数据底板构建技术研究[J].华中科技大学学报（自然科学版）,2026,54(7):48-54,7.基金项目
国家重点研发计划资助项目(2022YFC3002704) （2022YFC3002704）
湖北省水利重点科研项目(HBSLKY202412). （HBSLKY202412）
