the atmosphere at (d), and finally enters the
spacecraft instrument (e). It is this energy that
is measured by the instrument. In summary,
the quantity that we are seeking is the ratio of d
to c (the reflectance - which characterizes the
surface material), but the quantity measured is
e (the radiance entering the instrument). Part
of the processing that is undertaken must, by
one means or another, derive the reflectance
from the measured radiance taking into
account all of the factors which have been
noted above. We call this process reflectance
modelling.
Accurate modelling of the reflectance process
is quite a complex undertaking (Woodham &
Gray [1987]). The process is illustrated in
Figure 4. The energy entering the
instrument at point E has three components:
(1) energy reflected directly from
the target pixel (T), (2) energy
reflected from an adjacent pixel
(A) and scattered by the
atmosphere, and (3) energy
which is simply scattered by the
atmosphere. In addition, the
energy illuminating the target
pixel has several components:
direct illumination by the sun, light
from the sun which has been
scattered in the atmosphere, light
which has been reflected from
adjacent pixels and light which
has been reflected from adjacent
pixels and scattered by the
atmosphere. Depending on the
shape of the surface, and the
relative positions of the
illuminating source (the sun) and
the sensor, some of the the
pixels may not be directly
illuminated at all because they are
in the shadows of hills or
mountains. These pixels are
illuminated only by light which
has been scattered by the
atmosphere and/or reflected off
adjacent pixels. All of this is in
addition to the effects caused by
orientation of the target pixel. It is clear that in order to model this situation adequately, a digital
elevation model which describes the shape of the surface is necessary.
Figure 4 Reflectance Modelling
The Digital Elevation Model and its Importance
The above discussion has emphasized the importance of the availability of a digital elevation
model in accurately determining the reflectance of surface material. The digital elevation model is
critical in other respects as well. Radar backscatter is a strong function of the orientation of the
surface pixel relative to the incoming radiation. The ability to relate backscatter readings acquired