针对机载 Headwall 高光谱成像仪对水质监测能力的验证, 利用 2018 年 9 月 17 日航空高光谱数据和 17 个准同步实测水体浊度数据分别构建囫囵淖尔水体浊度波段比值反演模型、一阶微分反演模型和偏最小二乘反演模型, 估算同日的囫囵淖尔水体浊度的空间分布。研究结果表明: 利用机载 Headwall 高光谱数据构建的 3 个浊度反演模型, 验证均方根误差 RMSE 均小于验证样本浊度的极值差 5. 3 NTU, MRE 均小于 10% , 机载 Headwall 高光谱成像 仪能够较好地观测水体浊度差异; 在囫囵淖尔, 基于 Headwall 高光谱数据的偏最小二乘模型建模精度高于波段比 值模型和一阶微分模型, 决定系数 R2 达到 0.95, 综合误差 CE 为 1. 74% , 最适用于囫囵淖尔水体浊度的反演; 2018 年 9 月 17 日囫囵淖尔东部水域浊度范围为 21. 2~ 54.4 NTU, 呈现出北低南高的趋势, 湖中心区域水体浊度较低, 南部水域受水中藻类的影响, 水体浊度较高。利用航空高光谱遥感影像实现了浊度的定量反演, 为航空高光谱遥感数据用于水质参数反演提供借鉴。

To verify the ability of Headwall hyperspectral imagery for monitoring water quality , three turbidity retrieval models including band-ratio, first-derivative, and partial least-squares models were established with the Headw all airborne hyperspectral imagery and sy nchronous measured turbidities in Hulunnaoer on September 17th, 2018. The three models were adopted to estimate the spatial distributions of turbidity in Hulunnaoer. The results showed that: the three constructed models based on the airborne Headwall hy perspectral data verify that the root mean square error ( RMSE) is less than the verification sample turbidity extreme value differ ence of 5.3 NTU, and the MRE is less than 10% , the three models had good performances in turbidity retrieval and the Headw all hy perspectral imagery could be used for water quality retrieval; the partial least squares model with adeterm ination coefficient R2 ( 01 95) , and a comprehensive error CE( 1.74% ) showed the better performance compared to the band ratio model and the first-order differ ential model, which was the optimal model for turbidity retrieval in Hulunnaoer; the range of turbidity was between 21.2 and 54.4 NTU in the eastern area of Hunlunnaoer on September 17th, 2018, the distributions of turbidity in Hunlunnaoer showed an increasing trend from north to south, the turbidity in the center was relatively low , while the turbidity in the southern area was relatively high due to the exists of algae. This study quantitatively retrieved turbidity, which can provide a reference for remote sensing of water quality based on aerial hyperspect ral remote sensing imageries in the future.

:国家水体污染控制与治理科技重大专项( 2017ZX0710220012005); 国家重点研发计划( 2017YFC0405804; 2017YFC0405801) ; 兰州交通大学优秀平台支持( 201806) ; 国家重大科学研究计划( 2015CB953700) ; 民用航天“十三五”技术预先研究项目( Y7D0070038)