Optimization of spectral bands for ocean colour remote sensing of aquatic environments

Autor: Aneesh A. Lotliker, T. Srinivasa Kumar, K. H. Rao, Vinay Kumar Dadhwal, Ranganath R. Navalgund, T. Preethi Latha, P. V. Nagamani
Rok vydání: 2016
Předmět:
Zdroj: SPIE Proceedings.
ISSN: 0277-786X
DOI: 10.1117/12.2225870
Popis: Selection of central wavelengths, bandwidths and the number of spectral bands of any sensor to be flown on a remote sensing satellite is important to ensure discriminability of targets and adequate signal-to-noise ratio for the retrieval of parameters. In recent years, a large number of spectral measurements over a wide variety of water types in the Arabian Sea and the Bay of Bengal have been carried out through various ship cruises. It was felt pertinent to use this precious data set to arrive at meaningful selection of spectral bands and their bandwidths of the ocean colour sensor to be flown on the forthcoming Oceansat-3 of ISRO. According to IOOCG reports and studies by Lee and Carder (2002) it is better for a sensor to have ~15 bands in the 400-800 nm range for adequate derivation of major properties (phytoplankton biomass, colored dissolved organic matter, suspended sediments, and bottom properties) in both oceanic and coastal environments from observation of water color. In this study, ~417 hyper-spectral remote-sensing reflectance spectra (spectral range varies from ~380-800 nm) covering different water types like open, coastal, mid coastal and near coastal waters have been used to identify the suitable spectral bands for OCM-3. Central wavelengths were identified based on the results obtained from hyper-spectral underwater radiometer measurements of Rrs, HPLC pigments and spectrometer analyzed absorption spectra for all the above water types. Derivative analysis has been carried out from 1st to 5th order to identify the inflection and null points for better discrimination / identification of spectral peaks from the in situ Rrs spectra. The results showed that open ocean and coastal ocean waters has spectra peaks mostly in the blue, green region; turbid coastal waters has maximum spectral peaks in the red region. Apart from this, the spectral peaks were identified in the red region for the chlorophyll fluorescence in the open ocean and coastal waters. Based on these results 13 spectral bands were proposed in the VNIR region for the upcoming OCM-3 sensor. In order to obtain water leaving radiances from the measurements at spacecraft platform, it is necessary to do atmospheric correction we need to have spectral bands in the NIR and above regions. Hence, a set of bands 3 bands in the NIR and SWIR region were proposed for OCM-3 to address the atmospheric correction related issues.
Databáze: OpenAIRE