Conference Proceeding
Atmospheric correction matching to theoretical forest signature applied to different colombian regions
Author
Molina, Juan
Lugo, Pablo
Arias, Jeraldine
Guaje, Javier
Castro, Harold
Costa, Carlos
Cabrera, Edersson
Sanabria, Mayerling
Institutions
Abstract
The atmospheric correction of remote sensing products is a process frequently used to reduce the reflectivity error produced by the presence of aerosols and other particles in the atmosphere. This process applies linear transformations that converts digital values to values of reflectance. The 6S model is widely used to atmospherically correct the remote sensing products. This solar-radiation based model transforms the image to reduce the noise produced by the different atmospheric conditions. This model is a simulation-based method that requires an initial parameterization given by the at acquisition conditions. This parameterization is comprised by an atmospheric profile, aerosol profile, ground reflectance, geometries, and altitude of the image among others. In this paper, an automatic selection of the parameters using search of parameters is proposed to parameterize the model using as a reference the theoretical signature spectral of tropical forest on Landsat which is assumed to have a low dispersion. Our results show that the configuration with minimum distance is a atmosphere profile of mid-latitude winter, aerosol profile of biomass burning, ground reflectance of green vegetation and Aerosol Optical Thickness of 0.4.