Neural network modeling of surface chlorophyll and sediment content in inland water from Landsat Thematic Mapper imagery using multidate spectrometer data

Pranab Jyoti Baruah; Masayuki Tamura; Kazuo Oki; Hitoshi Nishimura

doi:10.1117/12.452815

14 January 2002 Neural network modeling of surface chlorophyll and sediment content in inland water from Landsat Thematic Mapper imagery using multidate spectrometer data

Pranab Jyoti Baruah, Masayuki Tamura, Kazuo Oki, Hitoshi Nishimura

Author Affiliations +

Proceedings Volume 4488, Ocean Optics: Remote Sensing and Underwater Imaging; (2002) https://doi.org/10.1117/12.452815
Event: International Symposium on Optical Science and Technology, 2001, San Diego, CA, United States

Abstract

Concentrations of chlorophyll and suspended sediment in surface water are tow important parameters for monitoring inland water quality. In the open ocean, it is not difficult to derive empirical algorithms relating received radiances at remote sensors to concentrations of water quality parameters. However, in optically complex inland water the task is difficult due to overwhelming of spectral signature of chlorophyll by other organic components present in high concentration. Neural Networks have been successful in modeling various geophysical transfer function. In this study, NN is used to model the transfer function between chlorophyll and sediment concentrations, and above-water upwelling reflectance simulated at three Landsat Thematic Mapper visible bands form spectrometer data. The developed model could estimate chlorophyll better than conventional regression analysis. In estimating surface chlorophyll, Root Mean Square Error (RMSE) for neural network was found to be < 15 percent, while the same for regression was > 30 percent. In estimating suspended sediment, regression performed comparatively better than in chlorophyll estimation with an RMSE of 22 percent. The corresponding RMSE for neural network was 12 percent. Upon validation, the trained model is used to get spatial distribution of the two water quality parameters from the Landsat Thematic Mapper imagery. Prior to this, the LandsatTM digital number values are converted to equivalent spectrometer-derived reflectances with a regression between these two quantities at sampling locations, thus taking into account the atmospheric effects which are often difficult to be satisfactorily quantified in inland waters.

Citation Download Citation

Pranab Jyoti Baruah, Masayuki Tamura, Kazuo Oki, and Hitoshi Nishimura "Neural network modeling of surface chlorophyll and sediment content in inland water from Landsat Thematic Mapper imagery using multidate spectrometer data", Proc. SPIE 4488, Ocean Optics: Remote Sensing and Underwater Imaging, (14 January 2002); https://doi.org/10.1117/12.452815

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available