AUTHOR=Huelsman Kelsey , Epstein Howard , Yang Xi , Mullori Lydia , Červená Lucie , Walker Roderick 

TITLE=Spectral variability in fine-scale drone-based imaging spectroscopy does not impede detection of target invasive plant species

JOURNAL=Frontiers in Remote Sensing

VOLUME=3

YEAR=2023

URL=https://www.frontiersin.org/journals/remote-sensing/articles/10.3389/frsen.2022.1085808

DOI=10.3389/frsen.2022.1085808

ISSN=2673-6187

ABSTRACT=<p>Land managers are making concerted efforts to control the spread of invasive plants, a task that demands extensive ecosystem monitoring, for which unoccupied aerial vehicles (UAVs or drones) are becoming increasingly popular. The high spatial resolution of unoccupied aerial vehicles imagery may positively or negatively affect plant species differentiation, as reflectance spectra of pixels may be highly variable when finely resolved. We assessed this impact on detection of invasive plant species <italic>Ailanthus altissima</italic> (tree of heaven) and <italic>Elaeagnus umbellata</italic> (autumn olive) using fine-resolution images collected in northwestern Virginia in June 2020 by a unoccupied aerial vehicles with a Headwall Hyperspec visible and near-infrared hyperspectral imager. Though <italic>E. umbellata</italic> had greater intraspecific variability relative to interspecific variability over more wavelengths than <italic>A. altissima</italic>, the classification accuracy was greater for <italic>E. umbellata</italic> (95%) than for <italic>A. altissima</italic> (66%). This suggests that spectral differences between species of interest and others are not necessarily obscured by intraspecific variability. Therefore, the use of unoccupied aerial vehicles-based spectroscopy for species identification may overcome reflectance variability in fine resolution imagery.</p>