Visualisation of spatial distribution of moisture content and basic density using near-infrared hyperspectral imaging method in sugi ( Cryptomeria japonica )

Abstract

This study quantified the spatial distribution of moisture content and basic density within Cryptomeria japonica branches using near-infrared (NIR) hyperspectral imaging and multivariate data analysis techniques. The hyperspectral data for wood disk samples were acquired using a Compovision NIR Composition Imaging System (Sumitomo Electric Industries, Ltd., Yokohama, Japan) with a spectral range of 913–2519 nm. The calibration models for prediction of moisture content and basic density were developed using partial least squares (PLS) regression analysis. The PLS model for moisture content and basic density showed good predictive performance with coefficients of determination 0·82 (RMSEP = 12·4%) and 0·84 (RMSEP = 28·0 kg m⁻³) respectively. The spatial distribution of predicted moisture content clearly detected the low moisture white zone surrounding the heartwood. The high density compression wood could be also visualised with the NIR hyperspectral imaging technique. The results presented here provide useful information for development of a novel log sorting method.