TY - JOUR
T1 - Spectral discrimination of Mediterranean maquis and phrygana vegetation: results from a case study in Greece
AU - Manevski, K
AU - Manakos, I
AU - Petropoulos, George
AU - Kalaitzidis, C
PY - 2012/4/3
Y1 - 2012/4/3
N2 - Mapping the spatial distribution of Mediterranean vegetation is crucial for understanding current ecosystem equilibrium and combating present phenomena, such as desertification and wildfires. Conclusive evidence on the spectral discrimination of such plants is thus necessary. To this end, this study focuses on the discrimination among three trees and three shrubs based on their spectral reflectance measured in a typical Mediterranean environment. Spectra from the plants were acquired by field spectroradiometry in the range between 350 and 2500 nm during an intensive field campaign that took place in Crete island in the spring 2010. Discrimination analysis was performed by applying non-parametric statistical tests on the unaltered spectral reflectance. The multivariate classificatory technique, employed for quantifying the shape similarity between the reflectance spectra, indicated that the majority of the plants possess distinct signatures from one another. The univariate tests implemented pointed out the existence of wavelengths where the plants can be discriminated. The use of unaltered reflectance narrows the statistical difference between the plants to bands in the visible and the shortwave infrared spectrum, but weakens the difference in the near-infrared spectrum, compared to continuum-removed reflectance analysis of the plants already published. The use of unaltered reflectance emphasizes detectable differences induced by the optical properties of the plants, as well as by variation of internal water of the plants related to drought adaptations. All in all, this work highlights the prospect of hyperspectral remote sensing in discriminating those plant species using field spectral libraries coinciding with high-quality radiometrically calibrated imagery.
AB - Mapping the spatial distribution of Mediterranean vegetation is crucial for understanding current ecosystem equilibrium and combating present phenomena, such as desertification and wildfires. Conclusive evidence on the spectral discrimination of such plants is thus necessary. To this end, this study focuses on the discrimination among three trees and three shrubs based on their spectral reflectance measured in a typical Mediterranean environment. Spectra from the plants were acquired by field spectroradiometry in the range between 350 and 2500 nm during an intensive field campaign that took place in Crete island in the spring 2010. Discrimination analysis was performed by applying non-parametric statistical tests on the unaltered spectral reflectance. The multivariate classificatory technique, employed for quantifying the shape similarity between the reflectance spectra, indicated that the majority of the plants possess distinct signatures from one another. The univariate tests implemented pointed out the existence of wavelengths where the plants can be discriminated. The use of unaltered reflectance narrows the statistical difference between the plants to bands in the visible and the shortwave infrared spectrum, but weakens the difference in the near-infrared spectrum, compared to continuum-removed reflectance analysis of the plants already published. The use of unaltered reflectance emphasizes detectable differences induced by the optical properties of the plants, as well as by variation of internal water of the plants related to drought adaptations. All in all, this work highlights the prospect of hyperspectral remote sensing in discriminating those plant species using field spectral libraries coinciding with high-quality radiometrically calibrated imagery.
UR - http://hdl.handle.net/2160/11406
U2 - 10.1109/JSTARS.2012.2190044
DO - 10.1109/JSTARS.2012.2190044
M3 - Article
SN - 1939-1404
VL - 5
SP - 604
EP - 616
JO - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
JF - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
IS - 2
ER -