Classificadores Maxver e Isodata na dinâmica de uso e ocupação do solo no pólo Petrolina-Pe/Juazeiro-Ba

Abstract

Throughout the national territory, the natural environment is impacted due to agricultural and urban expansion, causing numerous changes in land use and occupation, which implies the need for mapping and more in-depth studies of these areas. The Petrolina-PE/Juazeiro-BA hub is part of this expansion, and due to its socioeconomic importance, the objective is, through Geoprocessing and remote sensing techniques, to compare the MAXVER and ISODATA classifiers, associated with the study of landscape dynamics and the advance of irrigated agriculture over native vegetation. Methodologically, the data was analyzed using images from LANDSAT orbital platforms from the years 1975, 1990, 2001 and 2017. Image processing was carried out in the ArcGIS 10.2 software, and the choice of the best classifier was based on the Kappa index, calculated from the confusion matrix in the Excel 2013 software. After choosing the best classifier, temporal changes in land use in the region were evaluated using the cross tabulation. The two classifiers presented satisfactory results and are recommended for the study, however, the MAXVER classifier was the most recommended due to its higher Kappa index, higher percentages of agreement and lower errors of commission and omission. In the temporal analysis, it was possible to identify and quantify the changes in the landscape over the 42 years studied, with emphasis on the significant increase in agricultural and urban areas, and the reduction in areas of native vegetation of the arboreal caatinga type.