结合水体光学分类反演太湖总悬浮物浓度

本研究利用2008年11月、2009年4月、2010年5月及2010年8月的太湖水体原位观测数据,在对光学复杂水体进行光学分类的基础上,分别建立了针对各个类别水体的总悬浮物浓度高光谱反演模型.通过对每类水体中各个模型的性能比较,分别得到各类水体的最优模型:第一类水体,比值模型为最优模型;第二类水体,半分析模型2为最优模型;第三类水体,一阶微分模型为最优模型.同时,比较分类前后模型的精度和稳定性,结果表明分类后,两者均表现出不同程度的提高,并且分析了光学分类导致半分析模型精度下降的原因.最后针对本研究的结果在遥感数据上的适用性进行了探讨,表明在高光谱遥感数据上有很大的应用潜力.本研究结论对光学复杂湖泊水体的水色遥感具有积极重要的意义.

Hyperspectral Remote Sensing of Total Suspended Matter Concentrations in Lake Taihu Based on Water Optical Classification

Four field investigations were carried out in the Taihu Lake for collecting in situ observation data in Nov. 2008, Apr. 2009, May and Aug. 2010. On the basis of water optical classification, different retrieval algorithms were developed, specific for different types of waters. Based on the preformance of each model comparion in each type of waters, the optimal models obtained were 1 the band ratio model for Type 1 water; 2 the semi-analysis algorithm model 2 for Type 2 water; 3 the first-order differential model for Type 3 water. Meanwhile, an optimal retrieval model was also established using the same collection of calibration data. Some comparisons were done between the developed models for the classified and non-classified waters. The comparison results showed that the models for the classified waters had better performances than that for the non-classified water, in both the retrieval accuracy and the model stability. Then, analyses of the optical classification leading to the accuracy decrease of the semi-analysis algorithm model were processed. Finally, the results of this study in hyperspectral remote sensing date showed a great application potential by analysis. The findings of this study are significant for promoting the development of water color remote sensing for optically complex turbid inland waters.