A novel chlorophyll-a inversion model in turbid water for GF-5 satellite hyperspectral sensor——a case in Poyang Lake
DENG Shiquan1， TIAN Liqiao1， LI Jian2， CHEN Xiaoling1， SUN Zhaohua3， ZHANG Li4， WEI Anna1
1.State Key Laboratory of Information Engineering in Surveying， Mapping and Remote Sensing， Wuhan 430079， China；2.School of Remote Sensing and Information Engineering， Wuhan University， Wuhan 430079， China；3.Key Laboratory of Ocean and Marginal Sea Geology， South China Sea Institute of Oceanology， Chinese Academy of Sciences， Guangzhou 510301， China； 4.Key Laboratory of Poyang Lake Wetland and Watershed Rsearch， Ministry of Education， Jiangxi Normal University， Nanchang 330022， China
Abstract：In order to analyze the potential of GaoFen-5(GF-5) hyperspectral satellite data for water color Remote Sensing application in inland waters， 134 in situ spectral data sets by SVC HR1024 spectrometer， corresponding chlorophyll-a (Chl-a) concentration， and total suspended sediments (TSS) concentration were obtained from Poyang Lake in 2009， 2011， and 2016. Considering spectral response differences of Chl-a variation in high turbid water area， we utilized Normalized Difference Chlorophyll-a Index (NDCI) for optical partition. Then， the spectral simulation and difference processing were carried out for hyperspectral cross-derivative inversion of Chl-a concentration. Our results are that the optical partitioning is helpful to improve the inversion accuracy. The coefficient of determination (R2)， Root Mean Square Error (RMSE) and Mean Relative Error (MRE) of inversion model is 0.82， 1.12， and 0.32 in low-turbid water (NDCI≥0.06)， separately. However， the R2 of inversion model is only 0.37 in the whole lake area. Compared with the previous optimal multi-band inversion model (R2＝0.76) in Poyang Lake， our method is more effective which used original spectrum and difference spectrum (first difference， second difference) in the central wavelengths at 495， 591， 675， 679， 684， 688， 692， and 696 nm to establish a cross-derivative multi-band combination inversion model. However， further study is still needed for the hyperspectral inversion potential in Chl-a concentration in high turbid water (NDCI <0.06) with TSS concentration higher than 50 mg/ L.
邓实权,田礼乔,李建,陈晓玲,孙兆华,张琍,韦安娜. 面向GF-5卫星高光谱传感器的浑浊水体叶绿素a浓度反演算法研究——以鄱阳湖为例[J]. 华中师范大学学报(自然科学版), 2018, 52(3): 409-415.
DENG Shiquan,TIAN Liqiao,LI Jian,CHEN Xiaoling,SUN Zhaohua,ZHANG Li,WEI Anna. A novel chlorophyll-a inversion model in turbid water for GF-5 satellite hyperspectral sensor——a case in Poyang Lake. journal1, 2018, 52(3): 409-415.
2018, Vol. 52 
