Autoria: MELLO, M. P. de; PRADO, B. R. do; SILVA, G. B. S. da; SHIMABUKURO, Y. E.; SANO, E. E.

Resumo: ABSTRACT - The Cerrado biome covers more than two million km² in the central region of Brazil. It is the second largest biome in this country, presenting rich biodiversity. Remote sensing is a useful tool for characterizing land covers. However, because of its particularities and vegetation diversity, Cerrado is not easily distinguishable in optical images. So the use of remote sensing for discriminating Cerrado vegetation requires an investigation of the behavior of their spectral characteristics. This paper analyzes the behavior of vegetation cover in the Cerrado biome through the Landsat5/TM spectral bands and the corresponding synthetic bands derived from mixing models. It also seeks the ways to facilitate the discrimination of the vegetation type classes by defining specific indices using remote sensing tools. The Brasília National Park was chosen as the study area for this research. The methodology used the Linear Spectral Mixture Model applied to two sets of endmembers (green vegetation, bare soil and shade; and green vegetation, dry vegetation and shade) to generate the corresponding fraction images (synthetic bands), which could discriminate the vegetation classes of interest. The use of dry vegetation as one of the endmembers showed more promising results than the use of bare soil for distinguishing Cerrado?s phytophysiognomies.

Ano de publicação: 2009

Tipo de publicação: Artigo em anais e proceedings

Unidade: Embrapa Cerrados

Palavras-chave: Cerrado, Comportamento espectral, Fitofisionomia, Fitofisionomias do cerrado brasileiro, Landsat, Modelos Lineares de Mistura Espectral, Remote sensing, Sensoriamento remoto, Vegetation, Vegetação, Índices de bandas sintéticas