论文信息 - Water Constituents Inversion in Taihu Lake Based on Artificial Neural Network and Bio-optical Model

Water Constituents Inversion in Taihu Lake Based on Artificial Neural Network and Bio-optical Model

Inland water and coastal areas are usually called Case 2 water, characterized optically by high concentrations of colored suspended matter, various phytoplankton pigments and colored dissolved organic matter (CDOM), and inland water monitoring using remote sensing technique is still experimental, and its development depends on the improvement of remote sensors and inversion algorithms. This paper constructed a Bio-optical model in Taihu Lake based on the optical property of water active constituents, and then the Bio-optical model was used to create reflectance data sets corresponding to the central channel wavelengths of the channels of MODIS instrument in 400 nm- 700 nm, which are often considered in water constituents inversion. Lastly, the datasets which are created by Bio-optical model trained and constructed a NN model for water constituents inversion. The study showed that combination of Bio-optical model and NN technology is a very useful method for water quality monitoring in Taihu Lake.

[1] Roland Doerffer,et al. Neural network for emulation of an inverse model: operational derivation of Case II water properties from MERIS data , 1999 .

[2] M. Kishino,et al. Retrieval of Chlorophyll a, suspended solids, and colored dissolved organic matter in Tokyo Bay using ASTER data , 2005 .

[3] Hui Feng,et al. Modeling spectral reflectance of optically complex waters using bio-optical measurements from Tokyo Bay , 2005 .

[4] D. Doxaran,et al. Spectral signature of highly turbid waters: Application with SPOT data to quantify suspended particulate matter concentrations , 2002 .

[5] E. Fry,et al. Absorption spectrum (380-700 nm) of pure water. II. Integrating cavity measurements. , 1997, Applied optics.

[6] H. Claustre,et al. Variability in the chlorophyll‐specific absorption coefficients of natural phytoplankton: Analysis and parameterization , 1995 .

[7] C. Mobley,et al. Neural network training: Using untransformed or log‐transformed training data for the inversion of ocean colour spectra? , 2006 .