Identification of Pinus species related to historic architecture in Korea using NIR chemometric approaches

Abstract

In the present study, we identify similar Pinus species that have been used as building materials for traditional architecture in Korea. Discriminant models were designed to identify wood species of Pinus densiflora forma erecta Uyeki, P. densiflora Sieb. et Zucc., and P. sylvestris L. grown in Russia and Germany using near-infrared (NIR) spectroscopy coupled with multivariate analysis. Wood block samples are more practical to use in the field for wood identification; however, the face measurements have different spectral characteristics. Thus, we also prepared tablet samples from wood powder to determine the key factors for species identification without the spectral differences resulting from the wood faces. P. densiflora for. erecta and P. densiflora were clearly classified from P. sylvestris grown in Russia and Germany with a correct prediction rate of 100 % in both sample types. The discriminant models using wood block samples exhibited good performances as the Rp2 values corresponding to NIR spectral regions of 8000–4000 cm−1 were higher than 0.90. The discriminant models using tablet samples also showed good discriminant performance. The tablet samples reduced the spectral differences of each species in the second derivative NIR spectra. However, using the tablet samples provided more specific species information, resulting in a more accurate classification.