ANALYSIS OF THE EFFECTS OF ATMOSPHERIC CORRECTION ON ORBITAL IMAGES FOR STUDIES IN INTERIOR WATER BODIES

• Autor(a) principal: Segedi, Gabriella Correa
• Data de Publicação: 2023
• Outros Autores: Cicerelli, Rejane Ennes, Almeida, Tati, Roig, Henrique Llacer, Olivetti, Diogo, Bernardi, José Vicente Elias, Castreghini, Adriana
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Ra'e Ga (Online)
• Texto Completo: https://revistas.ufpr.br/raega/article/view/93465
• topic: Geociências, Sensoriamento Remoto, Hidrologia; Inland water bodies; Atmospheric correction; Landsat-8; Sentinel-2

The water reservoirs, in addition to their significance in electricity generation, serve as vital resources for various other requirements of the population. Images from orbital sensors have been applied to complement the monitoring of these environments and thus overcome the deficiency of spatial and temporal coverage of traditional techniques. However, studies involving water quality are still a great challenge due to the low signal coming from the water body and the interference of external factors (or environmental factors). Image correction/improvement procedures are often proposed, mainly to reduce atmospheric interference. In this study the best available atmospheric correction techniques were evaluated in order to indicate the technique that most closely matches the spectral response of remotely sensed images obtained in the field. During the study six atmospheric correction algorithms were applied (FLAASH, Second simulation of a Satellite Signal in the Solar Spectrum (6S), L8SR, Aquatic Reflectance (NASA/USGS), ACOLITE and Sen2Cor) that, based on the statistical analysis of discriminant analysis and covariance, indicated the 6S for Landsat and Sentinel images and ACOLITE for Landsat images as the most accurate. Although 6S showed a response close to the reference data, low variability in spectral response was observed. For time series, ACOLITE showed better capacity to correct the data. The type of application is also a preponderant factor, since it was evident that the use of time series indicated a different atmospheric correction technique when compared to the analysis of the scenes individually.