Determination of moisture content and density of fresh-sawn red oak lumber by near infrared spectroscopy

Rapid, nondestructive prediction of green moisture content (MC) and density can improve wood processing, for example by allowing presorting of lumber into different classes for optimal drying. Near infrared (NIR) spectroscopy, coupled with multivariate analytic statistical techniques, has been used to predict the MC and basic density of solid red oak (Quercus spp.) wood. Samples were prepared from fresh-sawn lumber purchased from a sawmill in east Tennessee. NIR spectra were collected from tangential, radial and transverse surfaces of the samples. Each property was correlated with spectra from 1000 to 2300 nm using projection to latent structures (PLS) models. PLS models were then validated using an independent test set. In general, spectra collected from transverse and radial surfaces gave better predictions than the ones collected from tangential surfaces. Good predictions were obtained for spectra collected from transverse or radial surfaces, with root mean square of errors of prediction (RMSEP) of less than 3.6 percent for MC and 19.8 kg/m 3 for basic density when using 8 principal components. NIR has the potential to be used for rapid in-line measurement of green MC and basic density of red oak lumber.