淀山湖叶绿素a的高光谱遥感监测研究

为实现淀山湖水体叶绿素a的快速监测， 于2006年春季对淀山湖进行了现场光谱测量和同步的采样、实验室分析，通过分析叶绿素a和光谱数据之间的关系，建立了多个基于高光谱数据的叶绿素a浓度的估算模型。结果表明：利用归一化反射率建立的模型中精度最高的是采用708 nm反射率的一元二次模型，R2为0.608;利用R710nm/R667nm反射比建立模型精度得到提高，R2接近0.68；利用695.5 nm一阶微分建立的指数模型精度进一步提高，其R2达到了0.76；分别利用708 nm和667 nm的归一化反射率对数作自变量进行多元回归得到的模型精度最高，R2接近0.8，可以用来实现对叶绿素a的定量估算，实现叶绿素a的快速监测。

APPLICATION OF HYPER-SPECTRAL REMOTE SENSING IN MONITORING CHLOROPHYLL-A CONCENTRATION IN LAKE DIANSHAN

In order to develop a fast way for monitoring chlorophyll-a concentration in Lake Dianshan of Shanghai,the reflection spectra in spring of 2006 was measured with simultaneous sampling and analysis,and the correlation between chlorophyll-a concentration and reflection spectra was studied on the basis of different chlorophyll-a algorithms.The accuracy of the linear algorithm using the normalized reflectance is not high(R~2<0.6).The relationship between chlorophyll-a concentration and the band-ratios R_(708)/R_(667) can improve the accuracy(R~2=0.68).The exponential algorithm between chlorophyll-a concentration and the first order differential of the reflectance at 695.5 nm is good(R~2=0.76).Multivariate regression models with the normalized reflectance at 708 nm and 607 nm and their logarithm were also established with a higher accuracy,and the best model was achieved with the logarithm at 708 nm and 607 nm(R~2=0.8).The present study thus demonstrated the potential of monitoring chlorophyll-a concentration by using hyper-spectral remote sensing technology in inland waterbody.