Identificação de áreas mineradas a partir de Sensoriamento Remoto: um olhar com o Mapbiomas

Abstract

Mining activity, often installed in hard-to-reach regions, is responsible for causing distant changes in land use and land cover. To monitor and identify changes the application of appropriate remote sensing tools is a viable alternative. There are different software and platforms if remote sensing that enable the digital processing of satellite images, Google Earth Engine (GEE) is a fast tool that brings the possibility of historical series analysis that help in sizing the scope of the impacts caused by mining activity. This work brings the comparison between some remote sensing methods that can be used to identify areas: SIGMINE, MAPBIOMAS, GLOBALFOREST WATCH ANDDELIMITATION BYREGIONOF INTEREST, presenting the most appropriate among them. Although MAPBIOMAS was considered the platform that presented better results among the others analyzed, the tool MAPBIOMAS, in the first months of the release of the new collection, presented several inconsistencies in some features. Mapbiomas was applied to design mines that extract gold, copper, iron, magnesite and talc in the state of Bahia in Brazil, analyzing the expansion of its areas of operation over the 36-year interval, between 1985 and 2020. The expansion of metals, gold, copper and iron mining is influenced by the commercialization value of these materials in the market; however, internal factors can also impact on an enterprise. Some inconsistencies found during identification, or not, of areas mined by Mapbiomas, is due to the methodology of filtering and image stabilization applied by the platform.