基于原位高光谱的极高悬浮泥沙浓度监测模型研究  

Investigation of Monitoring Models for Extremely High Suspended Sediment Concentration Based on In-Situ Hyperspectral Data

作　　者：侯峥 侯英卓 邢前国[2,3] 马永刚 明星 刘昕悦 HOU Zheng;HOU Yingzhuo;XING Qianguo;MA Yonggang;MING Xing;LIU Xinyue(Ningxia Hydrological and Water Resources Monitoring and Early Warning Center,Yinchuan 750000,China;Key Laboratory of Coastal Environmental Processes and Ecological Remediation,Yantai Institute of Coastal Zone Research,Chinese Academy of Sciences,Yantai 264003,China;University of Chinese Academy of Sciences,Beijing 100049,China;College of Physical Science and Technology,Hebei University,Baoding 071002,China)

机构地区：[1]宁夏回族自治区水文水资源监测预警中心,宁夏银川750000 [2]中国科学院烟台海岸带研究所海岸带环境过程与生态修复重点实验室,山东烟台264003 [3]中国科学院大学,北京100049 [4]河北大学物理科学与技术学院,河北保定071002

出　　处：《人民黄河》2025年第4期28-31,37,共5页Yellow River

基　　金：山东省创新技术引导计划项目(YDZX2024021);宁夏回族自治区重点研发计划项目(2023BEG02052);中国科学院重点仪器研发项目(YJKYYQ20170048)。

摘　　要：悬浮泥沙浓度(SSC)是重要的水质监测参数。以宁夏清水河王团段为研究对象,基于2022年8月26日至11月5日原位高光谱数据与SSC构建了适用于极高含沙量水体的SSC反演模型,选用单波段模型、波段差值模型、波段比值模型、二元线性模型4种经验模型,对比其精度。结果表明:针对极高含沙量水体,单波段模型拟合效果不佳,模型决定系数(R^(2))不足0.25。波段差值模型、波段比值模型、二元线性模型一定程度上可以降低噪声的影响,提高模型R^(2)。整体上,波段差值模型效果较好,基于650~720 nm波段反射率与560~700 nm波段反射率构成的波段差值模型的R^(2)大于0.40,其中R687-R685模型的R^(2)最大,为0.76。Suspended sediment concentration(SSC)is a crucial water quality monitoring parameter.Taking the Wangtuan section of Qingshui River in Ningxia as the study area,an SSC inversion model suitable for waters with extremely high suspended sediment concentration was developed based on insitu hyperspectral data and SSC measurements from August 26 to November 5,2022.Four empirical models were selected for comparison,including single-band model,band difference model,band ratio model and binary linear model.The results in-dicate that for waters with extremely high suspended sediment concentrations,the single-band model performs poorly,with a coefficient of determination(R^(2))of less than 0.25.The band difference,band ratio and binary linear models are able to reduce the impact of noise and improve the model’s R^(2).Overall,the band difference model performs the best,with an R^(2) greater than 0.40 for the model based on the reflectance difference between the 650-720 nm and 560-700 nm bands.Among which,the R687-R685 model achieves the highest R^(2) value of 0.76.

关 键 词：原位高光谱 悬浮泥沙浓度 反演模型 宁夏清水河王团段 

分 类 号：TV145.3[水利工程—水力学及河流动力学]