Hyperspectral near Infrared Imaging of Wooden Surfaces Performed Outdoors and Indoors

Abstract

Hyperspectral near infrared imaging has been applied in a field study of fungal growth on a variety of wood substrates exposed in an outdoor environment over a six-month period. This study was performed as a follow up to a hyperspectral examination study of fungal growth on wood surfaces in a laboratory setting. Hyperspectral measurements were carried out both outdoors and indoors in order to explore the influence of the different light conditions. Segmentation of the mould growth on the wood surfaces was carried out using principal component analysis, spectral angle mapper and partial least squares-discriminant analysis. Growth curves showing the fungal growth over time were obtained for all the samples from the measurements performed outdoors. However, there are some challenges connected to studies of wood surfaces due to structures caused by growth rings, knots and sometimes cracks. These wood properties will cause a great variation in the spectra from the wood and also cause natural variation in the fungal growth. Determining specific classes in a classification model such as partial least squares-discriminant analysis is proposed as a way to overcome these issues. Moreover, the wood substrates exposed in an outdoor environment will have a colour change due to photodegradation of lignin, wetting/leaching of the upper layer of the wood surface and growth of a variety of wood discolouring fungi. Hyperspectral technology is a promising technique to study wood properties and we plan to carry out a study to be able to separate and model the different effects on colour degradation on wood surfaces.