Geology Reference
In-Depth Information
Suitable for
surface
fire detection
10
6
3000
k
10
5
Suitable for
underground
fire detection
10
4
1000
k
10
3
500
k
300
k
10
2
10
1
1
0.1
0.2
0.5
1 2
Wavelength (
μ
m)
5 0 0 0
Figure 14.2.4. Planck
s emission-spectral curves for black bodies at 300, 500, 1000 and 3000 K. Note that as the
temperature of the blackbody rises, the peak of the emitted radiation shifts to shorter wavelengths. The grey zones
are spectral regions suitable for temperature-estimation studies of coal fires. Figure by Anupma Prakash; adapted
from Lillesand et al. 2007.
'
This relationship is very important in thermal remote sensing. From the plot of Planck
s law, it is clear that to study
targets close to about 300 K (ambient temperature of the Earth), the TIR region of the spectrum centered on 10
'
m
is well suited. However, as the target temperature rises and reaches about 1000 K, the shortwave infrared region
(centered on approximately 2
μ
m) is better suited. This implies that to study underground coal fires that tend to
raise the temperature of the overlying land surface only subtly (less than a degree to about 15°), the 8
μ
m
spectral region is ideal. For fires in surface coal seams and coal dumps that reach much higher temperatures, the
shortwave infrared region is better suited (Prakash et al., 1997; Prakash and Gupta, 1999; Zhang et al., 2004). In
large, intense-blazing coal fires, the burn-area temperatures can reach well over 1000 K, and such fires can also be
investigated by remote sensing in the visible and near-infrared regions.
-
14
μ
The discussion so far has been limited to spectral emissions from a hypothetical blackbody. Natural substances are,
however, not blackbodies. Natural targets tend not to absorb all energy that is incident on their surface and
subsequently, do not emit all of the energy they absorb. The ratio of energy emitted by a material at a given
wavelength and temperature to the energy emitted by a blackbody at the same wavelength and temperature is
referred to as its spectral emissivity,
ε
λ
. Being a ratio, spectral emissivity has no unit and its value can theoretically
range from 0 to 1. Most natural substances show an emissivity value ranging from 0.7 to 0.97. Table 14.2.1 gives
the emissivity values of some natural materials encountered in a typical coal-mining area.
For quantitative analysis in thermal remote sensing, emissivity plays a role in two places: (1) In estimating the
spectral radiance of a natural substance, the Planck
'
s law needs to be adapted to include spectral emissivity.
'
Planck
s law then takes the form
hc
2
2
B
λT
¼
ε
λ
ð
14
:
2
:
3
Þ
5
λ
ðe
hc=λK
B
TÞ
−
1
Þ
where
ε
λ
is the spectral emissivity of the target material and all other factors are the same as defined in equation
14.2.2 and (2) Satellite images record the amount of energy reaching the sensor in the form of quantized digital
numbers (DNs), which is directly related to the spectral radiance,
B
λT
, of the target. This relationship between the
DN value and the spectral radiance is well defined for all spectral bands of an individual satellite and often
provided in either the satellite-image metadata or the data-user handbook for a particular satellite. Once the spectral
radiance value is known and an emissivity value is assigned for the target, inverting Planck
'
s law, it is easy to
