以武汉市东湖为研究区域,利用同步的MODIS-Terra气溶胶光学厚度数据为输入参数,采用FLAASH模型对2010年3月11日HJ-1A/B卫星CCD影像进行大气校正处理,并利用多年实测数据建立叶绿素a浓度、悬浮泥沙浓度、黄色物质吸收系数三要素神经网络反演模型,对水色三要素进行反演。通过对反演结果与实测数据的对比分析可知,悬浮泥沙浓度、黄色物质吸收系数和叶绿素a浓度的平均相对误差分别为28.052%、17.628%和35.621%,表明HJ-1A/B卫星CCD传感器基本能满足II类水体水色要素的遥感监测需求。
The CCD sensors onboard HJ-1A/B satellite,launched on Sep.6,2008,have the superiorities of high spatial and temporal resolution in water environment monitoring for small lakes.Taking Donghu in Wuhan as an example,atmospheric correction of HJ-1A/B CCD imagery was carried out using fast line-of-sight atmospheric analysis of spectral hypercubes(FLAASH) model,in which aerosol optical depth was retrieved from synchronous MODIS-Terra data.Then an artificial neural network(ANN) algorithm for the retrieval of chlorophyll-a,suspended sediments and absorption coefficient of colored dissolved organic matter(CDOM) was established with in-situ data obtained in three consecutive years.The results show that,compared with the in-situ data,the mean relative errors of the retrieved suspended sediments,absorption coefficient of CDOM and chlorophyll-a are 28.052%,17.628% and 35.621% respectively,which can meet the requirement of monitoring water color elements in inland waters.