IDENTIFICATION OF WOOD SPECIES BY ACOUSTIC-RESONANCE SPECTROMETRY USING MULTIVARIATE SUBPOPULATION ANALYSIS

The identification of wood species remains an issue in restoration inv olving rare, old, or disguised wood parts. Precise restoration is requ ired in reconditioning the works of designers such as Frank Lloyd Wrig ht, and a quick, reliable, and nondestructive method of identification facilitates this restoration. Acoustic-resonance spectrometry (ARS) i s an analytical method using interferences in resonance signals across a range of frequencies. Combined with multivariate analysis technique s, ARS is a solution to the problem of identifying wood species. Subpo pulation detection analysis of samples of 26 different wood species ac hieved complete differentiation among species (p = 0.01). The number o f bootstrap replications of the spectral data has a significant effect on differentiation among the woods, as does the type of spectral filt ering prior to subpopulation analysis. Acoustic-resonance spectrometry outperforms near-IR spectrometry by a wide margin in identification o f the same wood species.