A generic platform for hyperspectral mapping of wood
The use of calibrated near-infrared (NIR) spectroscopy for predicting of a range of solid wood properties is described. The methods developed are applicable to large-scale nondestructive forest resource assessment and to tree breeding and silviculture programs. A series ofPinus radiata D. Don (radiata pine) samples were characterized in terms of density, longitudinal modulus of elasticity (E L), and microfibril angle (MFA). NIR spectra were obtained from the radial/longitudinal face of each sample and used to generate calibrations for the measured physical properties. The relations between laboratory-determined data and NIR fitted data were good in all cases, with coefficients of determination (R 2) ranging from 0.68 for 100/MFA to 0.94 for densitystrip. A good relation (R 2 = 0.83) was also obtained forE L estimated using data collected by SilviScan-2. The finding suggests that an NIR instrument could be calibrated to estimate theE L of increment cores based on SilviScan data. In view of the rapidly expanding range of applications for this technique, it is concluded that appropriately calibrated NIR spectroscopy could form the basis of a testing instrument capable of predicting a range of properties from a single spectrum obtained from the product or from the raw material.