摘要
[目的]淤地坝是黄土高原重要的水土保持工程措施,准确估算淤地坝淤积泥沙量对科学评价其水土保持效益和动态安全具有重要的意义。[方法]本研究提出一种基于无人机高清影像和三维曲面重构技术的淤地坝泥沙估算方法,首先以韭园沟小流域为例,基于无人机影像提取淤地坝地形曲面模型并随机选取30个空沟建立虚拟坝;在此基础上,采用二阶多项式、三阶多项式、平滑样条及二项指数4种方法模拟虚拟坝的纵向沟谷线和横向沟坡线特征并基于空沟地形数据验证方法的可靠性;其次,基于最优重构特征线构成的数字地形网络,采用Hutchinson算法进行空间插值得到淤地坝原始曲面;最后,以淤地坝重构曲面和淤地坝实际参数为参考,通过表面体积模型估算淤地坝的沉积泥沙量。[结果](1)4种方法中二项指数精度最高,其纵向和横向沟坡线模拟的均方根误差(RMSE)分别为1.305 m和2.958 m,且R^(2)分别>0.994,0.967。(2)利用Hutchinson空间插值方法得到的30个虚拟坝的模拟三维曲面,与真实曲面相比,模拟的RMSE为1.03 m,R^(2)>0.998;(3)基于该模型估算的30个虚拟坝的泥沙量平均误差率6.54%。[结论]研究提出的技术方法具有较高的精度和可靠性,可快速准确估算黄土高原地区单沟淤地坝沉积泥沙。
[Objective]Check dams are essential soil and water conservation engineering measures on the Loess Plateau.Accurate estimation of sediment deposition volume in check dams is important for the scientific evaluation of their soil-water conservation benefits and dynamic safety.[Methods]This study proposed a sediment estimation method for check dams based on high-resolution unmanned aerial vehicle(UAV)imagery and three-dimensional surface reconstruction technology.Taking the Jiuyuangou watershed as an example,UAV imagery was utilized to extract the terrain surface model of check dams,and 30 empty gullies were randomly selected to establish virtual dams.Four methods(second-order polynomial,third-order polynomial,smoothing spline,and binomial index)were employed to simulate the longitudinal valley lines and transverse gully slope lines of the virtual dams.The reliability of these methods was verified using empty gully topographic data.A digital terrain network based on the optimal reconstructed characteristic lines was established,and spatial interpolation was performed with the Hutchinson algorithm to generate the original surface of the check dams.Using the reconstructed check dam surface and actual dam parameters as references,the sediment deposition volume was estimated through 3D analyst model.[Results](1)among the four methods,the binomial index model achieved the highest accuracy,with root mean square errors(RMSE)of 1.305 m and 2.958 m for the simulated longitudinal and transverse gully slope lines,respectively.Additionally,R^(2)values exceeded 0.994 and 0.967,respectively.(2)The three-dimensional surfaces of 30 virtual dams simulated using the Hutchinson spatial interpolation method had an RMSE of 1.03 m compared to the actual terrain,with R^(2)exceeding 0.998.(3)The proposed model estimated sediment volume for the 30 virtual dams with an average error rate of 6.54%.[Conclusion]The proposed method demonstrates high accuracy and reliability,enabling rapid and accurate estimation of sediment deposition in single-gully check dams on the Loess Plateau.
作者
张晨
常潇笛
赵建林
汪珍亮
杨节
Zhang Chen;Chang Xiaodi;Zhao Jianlin;Wang Zhenliang;Yang Jie(School of Geological Engineering and Geomatics,Chang′an University,Xi′an 710054,China)
出处
《水土保持研究》
北大核心
2025年第5期48-58,共11页
Research of Soil and Water Conservation
基金
陕西林业科技创新重点专项(SXLK2023-02-15)
国家自然科学基金(41907048)
中央高校基本科研费专项资金(300102260206)。
关键词
淤地坝
三维曲面
无人机
泥沙体积估算
check dam
three-dimensional surface
unmanned aerial vehicle
sediment volume estimation
