Skip to main content
SciX 2025 Main banner

(Late) Invited Presentation

Chemometrics

CHEM03: Young Investigators in Chemometrics

Rapid Classification of Gymnosperm Wood Using Laser-Induced Breakdown Spectroscopy and Multivariate Analysis: A Partial Least Squares Discriminant Analysis Approach

Thursday, October 9, 2025
10:50 AM - 11:10 AM EDT
Location: Meeting Room #3

    Presenting Author(s)

  • FD

    Fernanda F. Delgado

    Chair
    University of Delaware
    Newark, Delaware, United States

    • Non-Presenting Author(s)

  • EM

    Erin McClure-Price

    US Forest Service International Programs Wood Identification and Screening Center
    Ashland, Oregon, United States

  • KE

    Kent Elliott

    US Forest Service International Programs Wood Identification and Screening Center
    Ashland, Oregon, United States

    • Project Lead(s)

  • KB

    Karl S. Booksh, PhD

    Professor
    University of Delaware
    Newark, Delaware, United States

    • Submitter(s)

  • FD

    Fernanda F. Delgado

    Chair
    University of Delaware
    Newark, Delaware, United States

In fifteen words or less, explain the significance of this contribution (Novel Aspect).: LIBS with PLS-DA, achieves near-90% agreement differentiating gymnosperm species, showing potential for rapid wood identification.

Abstract Text: Laser-Induced Breakdown Spectroscopy (LIBS) offers a rapid and minimally destructive means of characterizing wood, providing complex spectral data that can be leveraged for chemometric classification. In this study, we investigated the ability of LIBS combined with partial least squares–discriminant analysis (PLS-DA) to differentiate twelve gymnosperm species. Quercus alba was included as a control, measured throughout the dataset to verify instrument stability and to ensure no systematic drift occurred during acquisition.

Models were built with varying numbers of latent variables, and performance was assessed using cross-validation. The optimal model employed nine components and achieved a Cohen’s kappa of 0.89, reflecting strong overall classification performance. While major species were reliably distinguished, misclassifications were observed in classes with limited sample sizes, highlighting the need for more balanced datasets.

These findings demonstrate the potential of LIBS combined with chemometrics as a promising tool for rapid, species-level wood identification in forestry and timber analysis.