Análise de imagem polarimétrica TerraSAR-X para classificação de uso e cobertura da terra na porção sudoeste da Amazônia Brasileira / TerraSAR-X image for classification the land use and land cover in the southwest portion of the Brazilian Amazon

AUTOR(ES)
FONTE

IBICT - Instituto Brasileiro de Informação em Ciência e Tecnologia

DATA DE PUBLICAÇÃO

24/02/2011

RESUMO

The objective of this study is to analyze the potential use of images from satellite TerraSAR-X, at StripMap acquisition mode, to map and identify the land use and land cover (LULC) in the southwestern Brazilian Amazon. Classifiers based on statistical functions for maximum likelihood (ML) method and based on frequency-based contextual and neural network classification technique (Context) was used in images of amplitude in dual polarization (HH and VV), a synthesized image derivative of the covariance matrix <HH.VV*> and entropy (Entropy (H)), derived from the decomposition of targets by Cloude and Pottier. The results of these classifications were evaluated by the confusion matrix and \textit{Kappa} index. Physiognomic-structural and field observation points (GPS) served as reference for classification of six land cover classes and to evaluate the performance of the approaches used. It can be stated that all components [A$_{<HH>}$, A$_{<VV>}$, A$_{<HH.VV*>}$ e A$_{Entropy}$], at Context Classification performed better with an overall accuracy of 78,33\% and \textit{Kappa} 0,70. This allows us to conclude on the important role of TerraSAR-X images for the characterization of thematic landscape in the southwestern Brazilian Amazon, including new tools for monitoring human activities and its environmental impact, adding information to already existing traditional optical products obtained in areas of high cloud cover, as it occurs in tropical rain forests.

ASSUNTO(S)

mapeamento florestal radar classificação polarimétrica amazônia forest mapping radar polarimetric classification amazon (region)

ACESSO AO ARTIGO

Documentos Relacionados