Please use this identifier to cite or link to this item: http://www.alice.cnptia.embrapa.br/alice/handle/doc/902332
Research center of Embrapa/Collection: Embrapa Territorial - Artigo em anais de congresso (ALICE)
Date Issued: 2011
Type of Material: Artigo em anais de congresso (ALICE)
Authors: OLIVEIRA, D. P. de
PULIDO, J.
FRANCHESCHINI, M. H. D.
COSTA, P. A. da
ARRUDA, G. P. de
DEMATTÊ, J. A. M.
SOUZA FILHO, C. R. de
VICENTE, L. E.
Additional Information: DANIEL PONTES DE OLIVEIRA, ESALQ/USP - UFC; JAVIER PULIDO, ESALQ/USP - UFC; MARSTON HÉRACLES DOMINGUES FRANCHESCHINI, ESALQ/USP; PRISCILLA ALVES DA COSTA, ESALQ/USP; GUSTAVO PAIS DE ARRUDA, ESALQ/USP; JOSÉ ALEXANDRE MELO DEMATTÊ, ESALQ/USP; CARLOS ROBERTO DE SOUZA FILHO, IG/UNICAMP; LUIZ EDUARDO VICENTE, CNPM.
Title: Caracterização espectral de solos utilizando espectrorradiômetro em laboratório e imagem de satélite hiperespectral.
Publisher: In: SIMPÓSIO BRASILEIRO DE SENSORIAMENTO REMOTO, 15., 2011, Curitiba. Anais... São José dos Campos: INPE, 2011.
Pages: p. 8476-8483.
Language: pt_BR
Keywords: Processamento de imagens.
Description: Data obtained with hyperspectral remote sensors have the advantage of containing a great spectral resolution, offering more details about spectral behavior of a particular target. The use of these images show high potential to describe soil mineralogical attributes. The main objective of this study was to obtain the spectral and mineralogical attributes of soils using hyperspectral satellite imagery and with data acquired at ground level; evaluation of a supervised classification routine for determination of soils texture; and estimate clay using multivariate analysis. Soil samples were collected at a 0-20cm depth and spectral measurements, texture and mineralogy analysis were made. Using GIS software, image processing and statistical packages, the information obtained in the laboratory has been analyzed. The use of hyperspectral imagery enhanced the mineralogical characterization of the studied area. The maximum likelihood classification algorithm showed great skill in distinguishing between four textures class created with the aim of hyperspectral data. The statistical method PLSR provided a satisfactory prediction of clay and sand, using data collected in the laboratory, with high coefficients of determination and low error values (RMSE).
Thesagro: Geologia
Sensoriamento Remoto.
Data Created: 2011-10-04
Appears in Collections:Artigo em anais de congresso (CNPM)

Files in This Item:
File Description SizeFormat 
VicenteSBSR.pdf1,12 MBAdobe PDFThumbnail
View/Open

FacebookTwitterDeliciousLinkedInGoogle BookmarksMySpace