EM 1110-2-2907
1 October 2003
(6) Summary. Spectral radiance is the amount of energy received at the sensor per
time, per area, in the direction of the sensor (measured in steradian), and it is measured
per wavelength. The sensor therefore measures the fraction of reflectance for a given
area/time for every wavelength as well as the emitted. Reflected and emitted radiance is
calculated by the integration of energy over the reflected hemisphere resulting from dif-
fuse reflection (see http://rsd.gsfc.nasa.gov/goes/text/reflectance.pdf for details on this
complex calculation). Reflected radiance is orders of magnitude greater than emitted ra-
diance. The following paragraphs, therefore, focus on reflected radiance.
b. Spectral Reflectance Curves.
(1) Background.
(a) Remote sensing consists of making spectral measurements over space: how
much of what "color" of light is coming from what place on the ground. One thing that a
remote sensing applications scientist hopes for, but which is not always true, is that sur-
face features of interest will have different colors so that they will be distinct in remote
sensing data.
(b) A surface feature's color can be characterized by the percentage of incoming
electromagnetic energy (illumination) it reflects at each wavelength across the electro-
magnetic spectrum. This is its spectral reflectance curve or "spectral signature"; it is an
unchanging property of the material. For example, an object such as a leaf may reflect
3% of incoming blue light, 10% of green light and 3% of red light. The amount of light it
reflects depends on the amount and wavelength of incoming illumination, but the per-
cents are constant. Unfortunately, remote sensing instruments do not record reflectance
directly, rather radiance, which is the amount (not the percent) of electromagnetic energy
received in selected wavelength bands. A change in illumination, more or less intense sun
for instance, will change the radiance. Spectral signatures are often represented as plots
or graphs, with wavelength on the horizontal axis, and the reflectance on the vertical axis
(Figure 2-20 provides a spectral signature for snow).
(2) Important Reflectance Curves and Critical Spectral Regions. While there are
too many surface types to memorize all their spectral signatures, it is helpful to be famil-
iar with the basic spectral characteristics of green vegetation, soil, and water. This in turn
helps determine which regions of the spectrum are most important for distinguishing
these surface types.
(3) Spectral Reflectance of Green Vegetation. Reflectance of green vegetation
(Figure 2-21) is low in the visible portion of the spectrum owing to chlorophyll absorp-
tion, high in the near IR due to the cell structure of the plant, and lower again in the
shortwave IR due to water in the cells. Within the visible portion of the spectrum, there is
a local reflectance peak in the green (0.55 m) between the blue (0.45 m) and red (0.68
m) chlorophyll absorption valleys (Samson, 2000; Lillesand and Kiefer, 1994).
2-24
Previous Page
