CLASSIFICAÇÃO NÃO-SUPERVISIONADA DE IMAGENS RAPIDEYE NO MAPEAMENTO DA COBERTURA DAS TERRAS DO DELTA DO PARNAÍBA, PIAUÍ

The aim of this paper was the mapping of the land cover classes in an area of the Delta do Parnaíba, Piauí, NE, Brazil using the unsupervised classification method in RapidEye images. Through digital processing of images in a GIS environment, it was possible to map 12 classes of land cover. The results showed that the highest percentage of the study area is covered by fields with the presence of shrub vegetation and a predominance of pasture. Other classes (mobile dunes and sandy shoreline, undergrowth and exposed land) characterize a high level of environmental vulnerability and risk of erosion, illustrating the need for sustainable management techniques. Based on the techniques and evaluation criterias, the mapping indicated very good agreement, emphasizing the quality of the visual interpretation of the image and the classification method employed. AMORIM, J. V. A.; VALLADARES, G. S.; PORTELA, M. G. T. CLASSIFICAÇÃO NÃO SUPERVISIONADA DE IMAGENS RAPIDEYE NO MAPEAMENTO DA COBERTURA DAS TERRAS DO DELTA DO PARNAÍBA, PIAUÍ Geosaberes, Fortaleza, v. 12, p.88-106, 2021. 89