贵州百花湖叶绿素a高光谱监测模型研究

为对百花湖的叶绿素a浓度进行高光谱模型监测,以更高效和在更大范围内监测湖水富营养化状况,结合百花湖水体的具体情况,对湖区水体布置采样点进行监测,获取了2012年4月和12月的实测水体高光谱数据和叶绿素a浓度数据。通过对获取数据的处理和分析,提取了叶绿素a的敏感波段并建立了叶绿素a浓度与水体光谱反射率相关模型。分析结果发现:1百花湖叶绿素a的敏感波段为706,707,678 nm和679nm,且敏感波段的组合运算可提高光谱数据与叶绿素a浓度之间的相关显著性;2通过对模型精度的对比分析,发现用(706+707)/(678+679)波段组合所建立的线性模型对叶绿素a浓度的反演精度最高;3以4月份的数据建立的模型经12月份的数据验证,精度也较高,说明该模型在一定范围内不受时间限制,可用来对百花湖叶绿素a的浓度进行及时监测。

Study on hyperspectral monitoring model of chlorophyll a in Baihua Lake

To monitor chlorophyll a concentration of Baihua Lake in Guizhou Province by hyperspectral model and reflect the eutrophication state of the lake more effectively and in a more large scope, according to the concrete conditions of the lake, sampling points in the lake were set and the hyperspectral data and chlorophyll a data from April to December of 2012 were taken. Through data analysis, the sensitive spectra bands to chlorophyll were taken and the correlation model between chlorophyll a and spectrum reflection ratio was built. The analysis results show: (1) the sensitive spectrum to chlorophyll a of Baihua Lake are 706,707,678 nm and 679 nm, the composition calculation of sensitive spectrum bands can improve the correlation significance; (2) through model accuracy analysis, it is discovered that the model based on the remote sensing reflection ratio variable (706+707)/(678+679) is best for the concentration prediction of chlorophyll a (Chla); (3)the model established by April data and checked by December data is high in accuracy, showing that the model is not limited in a certain time scope and can be used to timely monitor chlorophyll a in Baihua Lake. The model provides a quick and effective means to monitor chlorophyll a in the lake.