Spectral signatures (Endmembers) some of forest species in the republic of Croatia

Abstract

All objects reflect, absorb or emit electromagnetic radiation depending on the composition, creating unique patterns called spectral signatures or endemembers. Pure spectral samples are defined in ideal field or laboratory conditions, where the spectrum of reflection is obtained using a spectroradiometer focused on one surface. According to researches, most spectral-pure samples refer to mineral exploration. Spectral signatures of vegetation, unlike minerals, are dynamic (in spectral, spatial and temporal resolution), considerably demanding for collecting and documenting, and need to be incorporated with caution in spectral libraries. There are several spectral libraries (bigger and smaller) that are organized by chapters which consist of samples that are adequately reviewed and documented to determine the spectrum quality. In this study, the spectral signatures for several species in Croatia were isolated: Oak (Quercus robur L.), Common Beech (Fagus sylvatica L.), Silver fir (Abies alba Mill.), Norway spruce (Picea abies L.), White-berried mistletoe (Viscum album L. ssp. Abietis (Weisb.)) and Yellow Mistletoe (Loranthus europaeus Jacq.). The purpose of the research was to establish a spectral library for future research into hyperspectral scanners for tree species detection. For collecting spectral signatures, the hyperspectral line scanner ImSpector V9 was used to capture the visible and near infrared spectrum (430-900 nm). In addition, the FODIS solar radiation sensor was used to obtain the average value of the solar insolation at the time of recording. Recording was performed under controlled conditions. Samples were placed on the circular base with the indicated division for every 45 degrees exactly in the center of the optical axis of the scanner and were rotated circularly. Spectral images were then processed in ImageJ software where data was extracted for further analysis. After calculation of the mean values by species, comparisons were made between species. The obtained results showed overlaps in the visible part of the spectrum. In the near infrared part of the spectrum they differentiate from one another, apropos the results show that there is a difference between the spectral curves of the samples. The research carried out defines sampling procedures and obtained spectral signatures for the investigated species. Spectral signatures have become part of the spectral library, and the most significant result of the research is the ability to detect the species on hyperspectral images.