@InProceedings{SanchesSouz:2014:CoReAn,
author = "Sanches, Ieda Del'Arco and Souza Filho, Carlos Roberto de",
affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE)} and Institute
of Geosciences, University of Campinas (UNICAMP), Brazil",
title = "Continuum Removal Analysis of the chlorophyll absorption feature
to detect plant stress induced by liquid hydrocarbon
contamination",
booktitle = "Proceedings...",
year = "2014",
organization = "International Geoscience and Remote Sensing Symposium, (IGARSS).",
publisher = "IEEE",
abstract = "This work explores the continuum removal (CR) technique to detect
plant stress in visible/near infrared wavelengths. The red edge of
the 680 nm chlorophyll absorption is a key feature in plant stress
studies (e.g., [1], [8], [7]). The CR transformation consists of
estimating the absorptions not due to the band of interest and
removing their effects [2]. The CR technique normalises the
reflectance spectrum and isolates absorption features to allow
comparison between absorption bands on a common basis [4]. The CR
method was initially used in geological remote sensing and was
first applied to studies of leaf biochemistry by [5]. [6]
demonstrated that the distinction between healthy and contaminated
plant samples was improved when analysing the continuum-removed
reflectance compared to reflectance and first derivative of
reflectance data. In that study, they applied the CR using fixed
continuum channels (e.g. 550-570 nm). Pre-stipulated wavelengths
(i.e. left and right channels) used to determine the continuum
line for the CR of a specific absorption feature can be adjusted
for each sample to better represent the feature (PRISM software,
[3]). In the present study, a time series of close range canopy
reflectance data of a grass plant species (Brachiaria brizantha)
grown in liquid hydrocarboncontaminated (diesel DSL; gasoline-GSL)
soil was acquired with a portable spectrometer (ASD FieldSpec® 3
Hi-Res). The parameters describing the chlorophyll 680 nm
absorption feature (continuum channels, depth, width, and area)
were derived using the CR applied to the spectra using fixed (FIX)
and adjusted (ADJ) continuum channels. Differences between CR-FIX
and CR-ADJ results are observed in Figure 1 for all parameters but
the feature centre (Table 1). Left and right channels of the
continuum line differs some 16 nm and 8 nm on average,
respectively, for the FIX and ADJ methods. In addition, the mean
depth, width (FWHM full width at half maximum) and area of the 680
nm feature yielded with the ADJ technique showed higher values.
The analysis of the parameters estimated for the 680 nm absorption
feature (CR-ADJ) for each of the contamination treatments,
indicates that plants stressed by DSL and GSL display mean values
of depth, width and area substantially lower than healthy plants
(CTR) (Table 2). The results imply that to better characterize an
absorption feature, the application of the CR technique using
adjusted channels is superior and should be favoured in the
analysis. Plant stress in brachiaria grass induced by the
contamination of soil with DSL and GSL can be detected with
spectral feature analysis focusing on the depth, width or area of
the 680 nm chlorophyll absorption feature.",
conference-location = "Quebec",
conference-year = "2014",
isbn = "9781479957750",
label = "lattes: 2456184661855977 1 SanchesSouz:2014:CoReAn",
language = "en",
targetfile = "1407sanches.pdf",
urlaccessdate = "25 abr. 2024"
}