Aplicação de técnicas estatística multivariadas na classificação de imagens orbitais, na região do baixo São Francisco (SE/AL).

• Autor(a) principal: Dompieri, Márcia Helena Galina
• Data de Publicação: 2017
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UFS
• Texto Completo: http://ri.ufs.br/jspui/handle/riufs/7164
• topic: Ciências atuariais; Rio São Francisco (SE/AL); Sensoriamento remoto; Sentinel2; Análise de componentes principais; Remote sensing; Sentinel2; Principal component analysis; OUTROS::CIENCIAS ATUARIAIS

The classification process of an orbital image involves the extraction of one or more features of interest because of the difference in reflectance between surface constituent materials, along the electromagnetic spectrum. The scenes are equivalent to twodimensional arrays. The main goal of this manuscript is the classification of orbital images from the MSI sensor aboard the Sentinel-2 satellite. The study area corresponds to the counties located in the region of the Lower São Francisco River region, in the states of Sergipe and Alagoas, by means of multivariate statistical techniques (Principal Components Analisys and K-means). In addition, a geospatial database was organized containing the main activities practiced in the area in 2015 – productions of rice, sugar cane, fish, cattle and sheep, which was spatialized in frequency intervals in order to help the process of class association. The results indicated the degraded pasture as the most representative (28%). Almost 20% of the area was identified as exposed soils in a few urban settlements, and about 15% of the continental surface water identified showed a high occurrence of suspended sediment, indicating a strong sedimentation of the water bodies.