20
1: Richard lucas, Aled Rowlands, Olaf Niemann, Ray Merton
404-1020 nm and 400-1050 nm regions respectively. The SASI operates only
in the SWIR (from 820-2500 nm). Sensors operating across the full spectral
range include the AVIRIS (224 wavebands) and HyMap (l26 wavebands). Each
of these sensors provides contiguous spectral coverage, including those regions where absorption by atmospheric water occurs, facilitating atmospheric
absorption effects to be calculated. More recently, sensors designed to observe
across the full reflected spectral range plus the TIR wavelength range have
become operational, including the DAIS-7915 (Ben-Dor et al. 2002).
The spectral resolution and sampling interval are critical in determining the
accuracy to which features in the radiation spectra can be measured. The spectral resolution of the sensor is defined as the narrowest bandwidth over which
radiation is recorded, and is measured based on the FWHM of the instrument
response. The spectral sampling interval is the distance (in units of wavelength)
between spectral measurement points. Sensors with spectral resolutions of
10 nm or smaller were designed largely for aquatic and atmospheric applications and include the CASI -series of instruments, whilst those with bandwidths
greater than 10 nm were designed primarily for terrestrial applications (Curran
1994) and include the AVIRIS (FWHM of 10 nm with bands spaced "-' 10 nm
apart) and HyMap (FWHM of 15-20 nm). Spectral sampling intervals in the
order of 2 to 3 nm magnitude are commonly provided by hyperspectral sensors. For many sensors, the band positions and widths are fixed although
some (e.g., the CASI-series and AISA) offer programmable bands depending
upon the application. The ASAS was one of the few sensors that recorded reflectance at varying viewing geometries, thereby allowing measurement of the
bi-directional reflectance characteristics of surfaces. Airborne sensors operating in the short wave infrared (SWIR) and TIR regions are listed in Table 1.2.
1.3.1.2
Spatial Resolution and Coverage
The spatial resolution of an observing sensor refers to a distance between the
nearest objects that can be resolved, is given in units oflength (e. g., meter) and
depends on the instantaneous field of view (IFOV). For many airborne sensors,
the spatial resolution is dictated largely by the flying height of the aircraft as
well as the configuration of the sensor and, in some cases, the aircraft platform
needs to be changed to achieve the required resolution (Vane et al. 1993). The
flying height of the aircraft also influences the width of the scan and hence the
extent of coverage. The lens optics and the integration time will similarly limit
the potential spatial resolution of the image, although data can be acquired at
spatial resolutions finer than 1 m.
The level of detail able to be resolved at different spatial resolutions is
indicated in Fig. l.la-d, which compares data acquired by several airborne
and also spaceborne sensors over an area of subtropical woodland in central
Queensland, Australia. Using aerial photography, tree crowns can be resolved
easily through differentiation between photosynthetic vegetation (PV), nonphotosynthetic vegetation (NPV; e.g., branches) and soil background, and
different tree species can be distinguished. Using 1 m spatial resolution CAS I
1: Richard lucas, Aled Rowlands, Olaf Niemann, Ray Merton
404-1020 nm and 400-1050 nm regions respectively. The SASI operates only
in the SWIR (from 820-2500 nm). Sensors operating across the full spectral
range include the AVIRIS (224 wavebands) and HyMap (l26 wavebands). Each
of these sensors provides contiguous spectral coverage, including those regions where absorption by atmospheric water occurs, facilitating atmospheric
absorption effects to be calculated. More recently, sensors designed to observe
across the full reflected spectral range plus the TIR wavelength range have
become operational, including the DAIS-7915 (Ben-Dor et al. 2002).
The spectral resolution and sampling interval are critical in determining the
accuracy to which features in the radiation spectra can be measured. The spectral resolution of the sensor is defined as the narrowest bandwidth over which
radiation is recorded, and is measured based on the FWHM of the instrument
response. The spectral sampling interval is the distance (in units of wavelength)
between spectral measurement points. Sensors with spectral resolutions of
10 nm or smaller were designed largely for aquatic and atmospheric applications and include the CASI -series of instruments, whilst those with bandwidths
greater than 10 nm were designed primarily for terrestrial applications (Curran
1994) and include the AVIRIS (FWHM of 10 nm with bands spaced "-' 10 nm
apart) and HyMap (FWHM of 15-20 nm). Spectral sampling intervals in the
order of 2 to 3 nm magnitude are commonly provided by hyperspectral sensors. For many sensors, the band positions and widths are fixed although
some (e.g., the CASI-series and AISA) offer programmable bands depending
upon the application. The ASAS was one of the few sensors that recorded reflectance at varying viewing geometries, thereby allowing measurement of the
bi-directional reflectance characteristics of surfaces. Airborne sensors operating in the short wave infrared (SWIR) and TIR regions are listed in Table 1.2.
1.3.1.2
Spatial Resolution and Coverage
The spatial resolution of an observing sensor refers to a distance between the
nearest objects that can be resolved, is given in units oflength (e. g., meter) and
depends on the instantaneous field of view (IFOV). For many airborne sensors,
the spatial resolution is dictated largely by the flying height of the aircraft as
well as the configuration of the sensor and, in some cases, the aircraft platform
needs to be changed to achieve the required resolution (Vane et al. 1993). The
flying height of the aircraft also influences the width of the scan and hence the
extent of coverage. The lens optics and the integration time will similarly limit
the potential spatial resolution of the image, although data can be acquired at
spatial resolutions finer than 1 m.
The level of detail able to be resolved at different spatial resolutions is
indicated in Fig. l.la-d, which compares data acquired by several airborne
and also spaceborne sensors over an area of subtropical woodland in central
Queensland, Australia. Using aerial photography, tree crowns can be resolved
easily through differentiation between photosynthetic vegetation (PV), nonphotosynthetic vegetation (NPV; e.g., branches) and soil background, and
different tree species can be distinguished. Using 1 m spatial resolution CAS I