Identification of Dead and Sound Knots by near Infrared Spectroscopy

Abstract

This paper examines the identification of dead and sound knots on the wood surface using near infrared (NIR) spectroscopy. Fifty dead and sound knots samples were prepared and NIR diffuse-reflectance spectra were collected from these samples. The spectra of normal wood were also obtained from the neighbourhood of each knot. The wavenumber range from 6600 cm−1 to 5400 cm−1 (1515 nm to 1852 nm) was employed for the multivariate analysis after taking into consideration characteristics of the spectra. The soft independent modelling of class analogies (SIMCA) classification technique was conducted for classification of dead knots, sound knots and normal wood classes. Specific spectral features were found in dead and sound knots at the following absorption bands: 6130 cm−1, 5980 cm−1, 5800 cm−1 and 5700 cm−1, which were related to bark tissue and compression wood. The SIMCA model could identify the differences between dead and sound knots as well as the presence of knots. The loadings and discrimination powers were consistent with the spectral features, showing peaks due to bark tissue and compression wood. The bands provide useful information to determine the effective range of electromagnetic waves for image analysis.