Discrimination of five commercial wood preservatives by handheld near-infrared spectrometers and Multivariate Data Analysis

IRG/WP 22-20686

M Rubini, P Dulucq, B Charrier

Nowadays, the recycling potential of wood waste is still limited, because of a lack of reliable lack of a reliable and fast device allowing the discrimination of preservative-treated wood that contains organic or/and inorganic contaminants. The purpose of this study was to set up a methodology, based on a SCiO low-cost handheld NIR spectrometer, and multivariate data analysis (MDA). Spectra was obtained on solid maritime pine (Pinus pinaster) wood impregnated with 5 different commercial preservatives at 3 different concentration levels (50, 70, 100%). Two classification methodologies were used: PLS-DA, and SVM-DA. Additional statistical analysis (Student's t-test with a significance level of 5%) highlight the absorption bands impacted by the presence of the wood preservative. For classification methodologies, several criteria were used to compare the performance of the classification models. Regarding the universal F1 Score performance criterion, SVM-DA model outperforms the PLS-DA model. Compared to PLS-DA model, SVM-DA model has been improved by 65%. Another criterion, such as, Sensitivity, indicated, For SVM-DA, that the correctly classified is comprised between 82 to 100%, whereas, for PLS-DA, the correctly classified is comprised between 43 to 90%. These findings show that the SCiO low-cost handheld NIR spectrometer coupled to MDA, such as SVM-DA model is useful to correctly classifying preservative-treated wood and opens new possibility in wood waste recycling.


Keywords: commercial preservatives, near-infrared, Pinus pinaster, recycling, wood

Conference: 22-05-29/06-02 Bled, Slovenia


Download document (439 kb)
free for the members of IRG. Available if purchased.

Purchase this document