SPECIES DIFFERENTIATION OF 2 COMMON LUMBER MIXES BY DIFFUSE-REFLECTANCE FOURIER-TRANSFORM INFRARED (DRIFT) SPECTROSCOPY

Diffuse reflectance Fourier transform infrared (DRIFT) spectroscopy wa s used to differentiate coniferous woods by species. Species mixtures studied were SPF (composed of white spruce [Picea glauca Voss], Engelm ann spruce [Picea engelmannii Parry], lodgepole pine [Pinus contorta D ougl.] and subalpine fir [Abies lasiocarpa Nutt.]) and HEM/FIR (compos ed of western hemlock [Tsuga heterophylla Sarg.], amabilis fir [Abies amabilis Dougl.] and Sitka spruce [Picea sitchensis Carr.]). DRIFT spe ctra of the green wood samples were used in SIMCA (simple modeling of class analogy) pattern recognition to differentiate species. This appr oach was able to classify SPF with 90% accuracy when all samples were identified, and 94% accuracy when poorly classified samples were elimi nated. For the HEM/FIR group, the method was able to classify all samp les with 81% accuracy, and up to 99% accuracy when poorly classified s amples were eliminated.