Wood colorimetry of Lauraceae species native to Brazil

Abstract

Considering the complexity and difficulty of identifying forest species, wooden disks were collected to verify the potential of colorimetry to distinguish native species from Araucaria Forest stands of the Lauraceae family. The following species were used: Nectandra megapotamica, Ocotea indecora, Ocotea diospyrifolia and Ocotea puberula. Nees, to provide data on these species that grow naturally in Santa Catarina state, southern Brazil, enriching a robust database that can be practically applied in the commercialization of native woods. Visible spectra and colorimetric parameters were obtained from each anatomical surface and the results were evaluated by comparing the mean of each species regarding radial trunk position and anatomical surface. The data were also submitted to principal component analysis and performance of discriminant models (k-NN, SVM and ANN) for species discrimination with raw and second-derivative data. In general, colorimetric data presented different behavior, and chromatic coordinates a* and b* had higher potential for distinguishing the species. According to the mean spectra, Ocotea indecora had reflectance values different from the other species. By principal component analysis, raw data indicated the separation only of Ocotea indecora, while second-derivative data allowed better distinction of species. In all discrimination models, second-derivative data produced the best results. Thus, the use of colorimetry has potential for wood distinction of the Lauraceae species evaluated, improving the oversight of illegally traded timber.