中文摘要：

本文依据2004年7月的实测数据构建了太湖夏季叶绿素a浓度的实测光谱数据估计模型,并使用2004年8月的数据对模型进行了验证。调查样点覆盖了太湖内的典型水域,水样数据由无锡太湖环境监测站采集。样点的光谱数据用ASDFieldSpec野外光谱仪获取,每个样点测量10次,测量结果被转换为遥感反射率。对不同的波段组合进行比较分析后,从可解释性出发,最终选择了归一化指数表达式作为最佳波段组合,所建立的模型为：Chla（μg/L）=EXP（2.478 ＋16.378＊N66）,其中,N66为（R696 -R661） /（R696 ＋R661）。模型的R^2为0.9051,显著性p〈0.0001。与其他模型相比,本文的模型比较稳健,用于估计8月的叶绿素a浓度具有较小的绝对误差。本文的工作同时表明,在太湖的夏季相邻月份,可以使用实测光谱数据模型进行水体叶绿素a浓度的估计。

英文摘要：

In this paper, a new hyperspectral data model that estimates chlorophyll-a concentration （Chla） in Taihu Lake of summer is proposed. The model was developed based on measurement in situ in July 2004 ,and was validated by hyperspectral data in August 2004. Water samples were collected by Wuxi Taihu Lake Environment Monitoring Station and covered the typical water areas. At each site, hyperspectral data were measured ten times by field spectroradiometer ASD FieldSpec, and were converted into remote sensing reflectance. Different band combinations were calculated and compared, and the normalized band index was selected because it is more explicable. The model built by data in July 2004 is Chla（μg/L） = EXP（2. 478 ＋ 16. 378 ＊ N66 ）, where N66 is （ R696 - R661 ）/（ R696 ＋ R661 ）. Goodness-of-fit statistics of the model R2 is 0. 9051 ,and p 〈0. 0001. Compared with other models,this one is more stable,and is of less absolute error when used to estimate Chla in August 2004. The works in the paper also showed that hyperspectral data model can be used to estimate Chla by month in the summer of Taihu Lake.