Image Processing Reference
In-Depth Information
Fig. 4.1
Spectral signal acquired by a remote sensor (l = wavelength, q = local incidence angle,
x/y = location on earth surface) (Schowengerdt
1997
)
upon the source function (sun), the extent to which the radia-
tion is modified by the atmosphere (downwelling and
upwelling atmospheric transmission) and the physical
and geometric structure and chemical constituents present
at the surface (reflectance). The sensor measures the
radiation in spectral bands, i.e., at a specific wavelength
or over a defined wavelength range (bandwidth). The
number of spectral bands, their bandwidths and locations
along the electromagnetic spectrum determine the spec-
tral capabilities or spectral resolution of sensors.
Most satellite sensors are multispectral systems like LANDSAT Thematic
Mapper (TM) or IKONOS. They sense the earth surface with a few broad spectral
bands. Sensors that are able to acquire a large number of spectral bands with narrow
bandwidths are called hyperspectral systems. They have high spectral sampling but
so far are limited to airborne or ground based systems except the only spaceborne
hyperspectral sensor HYPERION on the EO-1 satellite. Such detailed spectral
measurements, however, potentially allow for precise identification of the
chemical and physical material properties as well as surface geometry of surfaces (Clark
1999
). Related analyses are usually referred to as spectroscopy or imaging spectrometry.
spectral resolution
of a sensor is
determined by the
number of spec-
tral bands, their
bandwidths and
locations along the
electromagnetic
spectrum
