高光谱数据反演大气水汽研究

Retrieval of Water Vapor Using Hyperspectral Data

利用Hyperion高光谱数据,基于MODTRAN辐射传输模型进行模拟,分析表明943nm和953nm波段分别对干燥和湿润大气最敏感,是反演水汽的合适通道。提出针对Hyperion数据DCIBR方法反演大气总水汽含量,可以推广到环境与灾害监测预报小卫星星座超光谱数据的大气水汽反演。误差分析表明,大气散射作用以及地表反射率非线性变化,对DCIBR方法计算精度的影响比CIBR算法要小,大气模式引起的反演误差小于2%,通道的光谱特性是影响反演精度的重要因素。

Based on simulation of atmospheric transmittance of Hyperion data, 943nm and 953nm are sensitive to clry and humid condition respectively, and are suitable channels to retrieve water vapor（WV）. We present a Dual-CIBR（DCIBR）algorithm to retrieve WV for Hyperspectral Data. We find that DCIBR is less sensitive to atmospheric scattering and nonlinear variety Of reflectance, and error by different atmospheric model is less than 2%. Therefore, DICBR algorithm is suitable for Hyperion data and small satellite constellation Hyperspectral Data to retrieve WV.