Hyperspectral remote sensing applications for early stress detection of young plants

Remote sensing technologies have advanced significantly at last decades and have improved the capability to gather information about Earth's resources and environment. They have many applications in Earth observation, such as mapping and updating land-use and cover, deforestation, vegetation and water dynamics and quality, etc. In this study, the physical principles and some applications of two hyperspectral remote sensing techniques, reflectance and fluorescence, are briefly discussed with a view to achieve an early diagnosis of stress in young deciduous trees (Paulownia tomentosa) in response to adverse environmental conditions (abiotic stresses). Leaf reflectance and fluorescence data were collected in the visible and near infrared spectral ranges (350-1000 nm) using a portable fiber-optics spectrometer. Statistical analyses and spectral normalization procedures were used to account the changes in the spectral features of the trees in response of adverse conditions. Spectral analyses were performed at ten narrow bands in green, red, red edge and near-infrared spectral ranges. Fluorescence spectra were investigated at five characteristic wavelengths in a spectral region 600-850 nm. Spectral data analyses were compared with the results from the accompanying biochemical tests for the assessment of damage to the trees.