6533b873fe1ef96bd12d5473

RESEARCH PRODUCT

Simulation d'images d'un futur imageur multispectral géostationaire dédié à la couleur de l'eau : étude de l'influence des angles d'observation et d'éclairement sur la luminance mesurée et sur l'erreur d'estimation en chlorophylle

subject

description

Improvements of optical sensors now make it possible to launch medium spatial and high spectral resolution sensors on the Geostationary Earth Orbit (GEO). The GEO allows sensors to observe the Earth with high temporal resolution, unlike Low Earth Orbit (LEO) satellites. The objective of this thesis is to simulate radiance images that would be provided by a theoretical ocean color sensor, on a geostationary orbit, in order to characterize the images to be acquired, to specify the future instruments and to validate the image processing algorithms. For this purpose, we have combined a radiative transfer model for ocean waters (Hydrolight) and a radiative transfer model for the atmosphere (MODTRAN) to obtain simulated radiance images at the sea surface level and at the Top-Of-Atmosphere (TOA) level. The inputs of Hydrolight have been deduced from GlobColour Project data (chlorophyll, mineral suspended matter and organic matter) to simulate the Global Area Coverage (GAC) images of a geostationary sensor with low spatial resolution, medium spectral resolution and high temporal resolution. The analysis of these images enables to determine the radiance dynamic range for each spectral band over the day. The results indicate the angular limits to avoid the sunglint phenomenon and to ensure a minimum proportion of water leaving radiance in the measured TOA radiance. The simulated radiance images allow to evaluate the performance of an ocean color algorithm for chlorophyll concentration estimation in open waters. The results show that the error of chlorophyll estimation is moderately sensitive to observation and lighting angles, which are so important in geostationary images.