Analysis of landscape and local vegetation changes through the use of semi-automatic image classifiers: application in cases of environmental disasters. Dulcilene Edivirgem Nunes Procópio Câmara, Fabio Giordano, Laís Samira Correia Nunes

Abstract

The fire at the Ultracargo Petrochemical Terminal in the Port of Santos (SP) in 2015 caused severe environmental impacts, including atmospheric, water, and soil contamination. The analysis of land cover, particularly local vegetation affected by disasters, is important for identifying, quantifying, and monitoring changes in the landscape, as well as for impact management. In this context, the use of semi-automatic classifiers, which combine automated satellite image analysis with human intervention, has proven effective in tracking these changes. In this study, satellite images from the MapBiomas platform from 1985, 2014, 2015, and 2022 were used to analyze and compare changes in the landscape and local vegetation, through semi-automatic classification, in the area of the Port of Santos where the industrial accident occurred.