呼伦湖多波段遥感水深反演模型研究

为探求呼伦湖水深分布及水下地形情况，本文以5组Landsat TM/ETM+影像数据及湖区分布的41个实测水深点为基础，应用太阳光波段及热红外辐射波段组合模型进行水深模拟，通过对模型验证及误差分析，得出最优模型为对数模型；进一步的敏感性分析表明，以波段6的敏感度为最大。应用最优模型对波段进行运算，反演效果较好，以2007-09-24为最优年份。其结果证实了融入热红外辐射的多波段水深反演模型，适用于北方寒旱区水体并可推广应用。

Retrieval model of water depth in Hulun Lake using Multi-spectral Remote Sensing

In order to study the water depth and underwater terrain in the Hulun Lake, a composite model combining sunlight reflection band and thermal infrared radiation band has been developed based on the selected Landsat TM/ETM+ and data of water depth measured at 41 points, and applied to predict water depth in the lake. The model verification and errors analysis were also carried out. It is approved that the logarithmic model is the optimal one, and the largest sensitivity appears in band 6 in the model. Computed spectral data were used in logarithmic model, and the results show that retrieval water depth is reasonable and the most consistent with the real terrain occurred in Sept. 24, 2007. With this study it is confirmed that the model is suitable to the Cold and Arid Regions in north China.