Uso da espectroscopia na região do infravermelho próximo para predição de propriedades tecnológicas da madeira de eucalyptus

The knowledge of the wood properties has fundamental importance to indicate the potential and final use of this material. Non-destructive techniques are an alternative to characterize the wood in a fast, simple and reliable way. The near infrared spectroscopy (NIRS) has been used as a nondestructive...

ver descrição completa

Detalhes bibliográficos
Autor: Viana, Lívia Cássia
Formato: tesis de maestría
Estado:Versión publicada
Fecha de publicación:2008
País:Brasil
Recursos:Universidade Federal de Lavras (UFLA)
Repositorio:Repositório Institucional da UFLA
Idioma:portugués
OAI Identifier:oai:repositorio.ufla.br:1/2360
Acesso em linha:https://repositorio.ufla.br/handle/1/2360
Access Level:acceso abierto
Palavra-chave:CNPQ_NÃO_INFORMADO
Clone
Madeira
Eucalyptus
Wood
Descrição
Resumo:The knowledge of the wood properties has fundamental importance to indicate the potential and final use of this material. Non-destructive techniques are an alternative to characterize the wood in a fast, simple and reliable way. The near infrared spectroscopy (NIRS) has been used as a nondestructive method for the characterization of wood. This technique allows to acquire qualitative and quantitative information using the interaction of near-infrared electromagnetic waves with the constituents of the samples. This research aimed to apply the Near Infrared Spectroscopy Technique for fast prediction of the basic density, anatomical and chemical properties of the wood of Eucalyptus clones. Six 3-years old Eucalyptus clones from plantations located in Cocais, Guanhães, Rio Doce and Santa Barbara sites, from CENIBRA Company, in the Minas Gerais state were used in this research. The basic density, fiber dimensions and chemical characteristics of the wood were determined by the standard method in the Wood Science and Technology Laboratory, in the Universidade Federal de Lavras, Lavras - Minas Gerais. The spectra were acquired using two equipments: the near infrared spectrometers Bruker at the V & M Company, located in Paraopeba - Minas Gerais and the Femto spectrometer at the Wood Science and Technology Laboratory the Universidade Federal de Lavras. The values of the wood properties were correlated with the spectral information by the partial least square regression (PLS regression). The calibrations to estimate the wood density showed correlation coefficients in cross validation (Rcv) ranging from 0.33 to 0.95 and the ratio of performance to deviation (RPD) ranging from 1.2 to 4.3. The wood chemical properties were predicted by models of Rcv ranging from 0.42 to 0.96 and RPD from 0.8 to 3.6. The calibration for prediction of fiber dimensions showed Rcv ranging 0.36 and 0.97 and RPD from 1.0 and 3.8. The near infrared spectroscopy has proved efficiency to predict the wood properties of Eucalyptus clones.