Neural network for emulation of an inverse model - operational derivation of Case II water properties from MERIS data

被引:227
作者
Schiller, H [1 ]
Doerffer, R [1 ]
机构
[1] GKSS Forschungszentrum Geesthacht GmbH, D-21502 Geesthacht, Germany
关键词
D O I
10.1080/014311699212443
中图分类号
TP7 [遥感技术];
学科分类号
081102 ; 0816 ; 081602 ; 083002 ; 1404 ;
摘要
An algorithm was developed to derive the concentrations of phytoplankton pigment, suspended matter and gelbstoff, and the aerosol path radiance from 'Rayleigh corrected' top-of-atmosphere reflectances over turbid coastal waters. The procedure is designed for MERIS, the Medium Resolution Imaging Spectrometer, which will be flown onboard the Earth observation satellite Envisat of the European Space Agency (ESA). The algorithm is a neural network (NN) which is used to parameterize the inverse of a radiative transfer model. It is used in this study as a multiple nonlinear regression technique. The NN is a feedforward backpropagation model with two hidden layers. The NN was trained with computed reflectances covering the range of 0.5-50 mu g1(-1) phytoplankton pigment, 1-100mg1(-1) suspended matter, gelbstoff absorption at 420 nm of 0.02-2m(-1) and a horizontal visibility of 2-50 km. Inputs to the NN are the reflectances of the 16 spectral channels which were under discussion for MERIS. The outputs are the three water constituent concentrations and the aerosol concentration, here expressed as the horizontal ground visibility. Tests with simulated reflectances show: (1) that concentations are correctly retrieved for a wide range covering oligotrophic Case I and turbid Case II water; (2) that the atmospheric correction can be performed even over very turbid water where the reflectance of the water cannot be neglected for the atmospheric correction channels in the near-infrared spectral range; and (3) that the algorithm is robust against errors in the input data. Although the training of the NN is time consuming, the utilization of the NN algorithm is extremely fast and can be applied routinely for satellite data mass production.
引用
收藏
页码:1735 / 1746
页数:12
相关论文
共 14 条
[1]   CONJUGATE-GRADIENT NEURAL NETWORKS IN CLASSIFICATION OF MULTISOURCE AND VERY-HIGH-DIMENSIONAL REMOTE-SENSING DATA [J].
BENEDIKTSSON, JA ;
SWAIN, PH ;
ERSOY, OK .
INTERNATIONAL JOURNAL OF REMOTE SENSING, 1993, 14 (15) :2883-2903
[2]  
BUCKTON D, 1999, USE NEURAL NETS ESTI
[3]  
Cherkassky V., 1994, From Statistics to Neural Networks
[4]  
DOERFER R, 1994, J GEOPHYS RES, P7457
[5]  
DOERFFER R, 1992, EURO COURS REM SENS, V2, P215
[6]  
Esaias W.E., 1986, EOS Trans. Am. Geo. Union, V67, P835, DOI [10.1029/EO067i044p00835, DOI 10.1029/EO067I044P00835]
[7]  
Fischer J., 1987, ADV SPACE RES, V7, P21, DOI [DOI 10.1016/0273-1177(87)90159-1, 10.1016/0273-1177(87)90159-1]
[8]  
Gordon H. R., 1983, REMOTE ASSESSMENT OC
[9]   SUBSURFACE WATER PARAMETERS - OPTIMIZATION APPROACH TO THEIR DETERMINATION FROM REMOTELY SENSED WATER COLOR DATA [J].
JAIN, SC ;
MILLER, JR .
APPLIED OPTICS, 1976, 15 (04) :886-890
[10]  
KEALY P, 1999, INT J REMOTE SENS, V20, P1703