Near-infrared spectroscopy and wavelength selection for estimating basic density in Mimosa tenuiflora [Willd.] Poiret wood

Abstract

The aim of the present work was to estimate the basic wood density of Mimosa tenuiflora by using near-infrared spectroscopy (NIRS). Fifty-eight wood samples representing sapwood, heartwood and pith were evaluated by gravimetric method and NIRS together with wavelength selection methods. A comparison was made among several multivariate calibration techniques and algorithms for preprocessing and variable selection of data, including full-spectrum partial least squares (PLS), interval PLS, backward interval PLS, synergy interval PLS, genetic algorithm-PLS and successive projections algorithm for interval partial least squares (iSPA–PLS). Finally, the results obtained using iSPA–PLS model for the root mean square error of calibration and prediction were 0.0383 and 0.0166 g/cm3, respectively. A t-test was performed to compare the results of the models with each other and with a reference method. NIRS and iSPA–PLS can be used to predict basic density of Mimosa tenuiflora [Willd.] Poiret wood samples rapidly. In addition, the basic density could also be predicted with only 17 wavelengths in the range from 2,090 to 2,208 nm that should allow for measurement of this parameter using handheld NIR spectrometer.