Hyperspectral Sensors and Applications
33
of anthropogenic activity (Yentsch et al. 2002) and changes in regional and
global climate and particularly the EI Niiio Southern Oscillation (ENSO). Similarly, recognition that the sensitivity of corals to temperature (Riegl 2002),
light (Sakami 2000), salinity (Ferrier-Pages et al. 1999), turbidity (Perry 2003)
and oxygen availability actually rendered them as suitable "environmental
barometers" .
Hyperspectral sensors have enabled new avenues of coral reef research,
particularly in relation to the critical differentiation between healthy and environmentally stressed coral (Hochberg et al. 2003). Furthermore, some of the
difficulties associated with the use of conventional satellite sensors and even
IKONOS data (e. g., coarse spatial or spectral resolution) for discriminating
substrate features and principal geomorphological zones associated with coral
reefs (Maeder et al. 2002; Mumby and Edwards 2002) have also been partly
overcome using hyperspectral data. As with many applications, monitoring of
coral reefs using remote sensing data has provided a cost -effective and timeefficient tool for reef surveys, change detection, and management (Maeder et
al. 2002).
1.6.2
Vegetation
Hyperspectral sensors are well suited for vegetation studies as reflectance/absorption spectral signatures from individual species as well as more complex
mixed-pixel community scale spectra can be utilised as a powerful diagnostic
tool for vegetation sciences. The spectral geometry of vegetation signatures
vary according to, for example, biochemical content and the physical structure
of plant tissues, and is further influenced by phenologic factors and natural and
anthropogenically induced factors creating a complex matrix of influencing
variables. To "unscramble" the weighted influence of these variables upon the
resultant vegetation spectral signature remains the perennial challenge to the
monitoring of vegetation with hyperspectral systems.
1.6.2.1
Reflectance Characteristics of Vegetation
Knowledge of the causes of variation in the spectral reflectance of vegetation (Fig. 1.4) has been fundamental to understanding the information
content of spectra derived from ground-based or airborne/spaceborne spectrometers. In the visible (red, green, and blue) wavelength regions, plant reflectance is dictated by the amount and concentration of photosynthetic pigments, namely chlorophyll a, chlorophyll b, xanthophylls, anthocyanins and
carotenoids (Guyot et al. 1989; Cochrane 2000; Chisholm 2001). In the NIR
wavelength region, the internal structure of leaves and, in particular, the size,
shape and distribution of air spaces and also the abundance of air-water interfaces (and hence refractive index discontinuities) within the mesophylliayers
exert the greatest influence on reflectance (Knox 1997). Much of the radiation
33
of anthropogenic activity (Yentsch et al. 2002) and changes in regional and
global climate and particularly the EI Niiio Southern Oscillation (ENSO). Similarly, recognition that the sensitivity of corals to temperature (Riegl 2002),
light (Sakami 2000), salinity (Ferrier-Pages et al. 1999), turbidity (Perry 2003)
and oxygen availability actually rendered them as suitable "environmental
barometers" .
Hyperspectral sensors have enabled new avenues of coral reef research,
particularly in relation to the critical differentiation between healthy and environmentally stressed coral (Hochberg et al. 2003). Furthermore, some of the
difficulties associated with the use of conventional satellite sensors and even
IKONOS data (e. g., coarse spatial or spectral resolution) for discriminating
substrate features and principal geomorphological zones associated with coral
reefs (Maeder et al. 2002; Mumby and Edwards 2002) have also been partly
overcome using hyperspectral data. As with many applications, monitoring of
coral reefs using remote sensing data has provided a cost -effective and timeefficient tool for reef surveys, change detection, and management (Maeder et
al. 2002).
1.6.2
Vegetation
Hyperspectral sensors are well suited for vegetation studies as reflectance/absorption spectral signatures from individual species as well as more complex
mixed-pixel community scale spectra can be utilised as a powerful diagnostic
tool for vegetation sciences. The spectral geometry of vegetation signatures
vary according to, for example, biochemical content and the physical structure
of plant tissues, and is further influenced by phenologic factors and natural and
anthropogenically induced factors creating a complex matrix of influencing
variables. To "unscramble" the weighted influence of these variables upon the
resultant vegetation spectral signature remains the perennial challenge to the
monitoring of vegetation with hyperspectral systems.
1.6.2.1
Reflectance Characteristics of Vegetation
Knowledge of the causes of variation in the spectral reflectance of vegetation (Fig. 1.4) has been fundamental to understanding the information
content of spectra derived from ground-based or airborne/spaceborne spectrometers. In the visible (red, green, and blue) wavelength regions, plant reflectance is dictated by the amount and concentration of photosynthetic pigments, namely chlorophyll a, chlorophyll b, xanthophylls, anthocyanins and
carotenoids (Guyot et al. 1989; Cochrane 2000; Chisholm 2001). In the NIR
wavelength region, the internal structure of leaves and, in particular, the size,
shape and distribution of air spaces and also the abundance of air-water interfaces (and hence refractive index discontinuities) within the mesophylliayers
exert the greatest influence on reflectance (Knox 1997). Much of the radiation