Transcript Selected Applications of Thermal Infrared Remote Sensing Jensen, 2000, 2006

Selected Applications
of Thermal Infrared
Remote Sensing
Jensen, 2000, 2006
Nighttime Thermal Infrared Imagery of an Airport
Jensen, 2000, 2006
Thermal Infrared Remote Sensing
Thermal infrared energy is emitted from all objects that have a
temperature greater than absolute zero. Therefore, all features we
encounter in the landscape on a typical day (Sun, vegetation, soil, rocks,
water, people) emit thermal infrared electromagnetic radiation.
• Humans sense thermal energy primarily through the sense of touch.
Our eyes cannot detect differences in thermal infrared energy because
they are primarily sensitive to short wavelength visible light from 0.4
m to 0.7 m. Our eyes are not sensitive to the reflective infrared (0.7
- 3.0 m) or thermal infrared energy (3.0 - 14 m).
• Engineers have developed detectors that are sensitive to thermal
infrared radiation. These thermal infrared sensors allow humans to
sense a previously invisible world as they monitor the thermal
characteristics of the landscape.
Fundamental Properties of
Electromagnetic Radiation
The three basic ways in which energy can be transferred:
• Conduction occurs when one body (molecule or atom) transfers its
kinetic energy to another by colliding with it. This is how a pan is heated
on a stove.
• Convection occurs when the kinetic energy of bodies is transferred from
one place to another by physically moving the bodies. An example is the
convectional heating of air in the atmosphere in the early afternoon.
• The transfer of energy by electromagnetic radiation is of primary interest
to remote sensing because it is the only form of energy transfer that can
take place in a vacuum such as the region between the Sun and the Earth.
How is Energy Transferred?
Energy may be transferred three ways: conduction, convection, and radiation. a) Energy
may be conducted directly from one object to another as when a pan is in direct physical
contact with a hot burner. b) The Sun bathes the Earth’s surface with radiant energy
causing the air near the ground to increase in temperature. The less dense air rises,
creating
Jensen
2005 convectional currents in the atmosphere. c) Electromagnetic energy in the form
of electromagnetic waves may be transmitted through the vacuum of space from the Sun
to the Earth.
Earth Observing System - Terra Instruments
ASTER - Advanced Spaceborne Thermal Emission and Reflection Radiometer
Spectral Range
VNIR 0.52 – 0.86 m, SWIR 1.6 - 2.43 m, TIR 8 - 12 m
Spatial Resolution 15 m (VNIR : 3 bands)
30 m (SWIR: 6 bands)
90 m (TIR: 5 bands)
Jensen, 2000
Thermal Infrared Radiation Principles
• An analyst cannot interpret a thermal infrared image as if it were an
aerial photograph or a normal image produced by a multispectral
scanner or charge-coupled device.
• Rather, the image analyst must think thermally.
• The analyst must understand how energy from the Sun or from the
Earth interacts with the various terrain components and how the
detectors function as they record the terrain’s emitted thermal infrared
electromagnetic radiation. Finally, the analyst must understand how
both the sensor system itself and the terrain can introduce noise into the
thermal infrared image that might make the data less useful or lead to
incorrect image interpretation.
Characteristics of a
Thermal Infrared
Airborne Across-track
Scanner
Kinetic Heat, Temperature, Radiant
Energy and Radiant Flux
• The energy of particles of matter in random motion is called kinetic heat
(also referred to as internal, real, or true heat). All objects having a
temperature above absolute zero (0 K; -273.16 ˚C; and -459.69 ˚F) exhibit
this random motion. When these particles collide they change their energy
state and emit electromagnetic radiation as previously discussed.
• The amount of heat can be measured in calories (the amount of heat
required to raise the temperature of 1 g of water 1 ˚C). We can measure
the true kinetic temperature (Tkin) or concentration of this heat using a
thermometer. We perform this in situ (in place) temperature measurement
when we are ill. We can also measure the true kinetic internal temperature
of soil or water by physically touching them with a thermometer.
Kinetic Heat, Temperature, Radiant
Energy and Radiant Flux
• Fortunately for us, an object’s internal kinetic heat is also converted to
radiant energy (often called external or apparent energy). The
electromagnetic radiation exiting an object is called radiant flux () and
is measured in watts. The concentration of the amount of radiant flux
exiting (emitted from) an object is its radiant temperature (Trad).
• There is usually a high positive correlation between the true kinetic
temperature of an object (Tkin) and the amount of radiant flux radiated
from the object (Trad). Therefore, we can utilize radiometers placed some
distance from the object to measure its radiant temperature which
hopefully correlates well with the object’s true kinetic temperature. This
is the basis of thermal infrared remote sensing.
Kinetic Heat, Temperature,
Radiant Energy and Radiant Flux
Unfortunately, the relationship is not perfect, with the remote
measurement of the radiant temperature always being slightly
less than the true kinetic temperature of the object. This is due
to a thermal property called emissivity – to be discussed shortly.
Thermal Infrared Atmospheric Windows
• Beyond the visible region of the electromagnetic spectrum, we
encounter the reflective infrared region from 0.7 - 3.0 m and the
thermal infrared region from 3 - 14 m.
• The only reason we can use remote sensing devices to detect infrared
energy in these regions is because the atmosphere allows a portion of
the infrared energy to be transmitted from the terrain to the detectors.
Regions that pass energy are called atmospheric windows. Regions
that absorb most of the infrared energy are called absorption bands.
Water vapor (H2O), carbon dioxide (CO2), and ozone (O3) are
responsible for most of the absorption. For example, atmospheric water
vapor (H2O) absorbs most of the energy exiting the terrain in the region
from 5 to 7 m making it almost useless for remote sensing.
Atmospheric Windows in the Electromagnetic Spectrum
Reflective Infrared Detectors
• Remote sensors can be engineered to be sensitive to the infrared
energy present within the reflective infrared atmospheric windows.
• Film emulsions can be made sensitive to reflected infrared energy in
the window from 0.7 - 0.9 m. For example, Kodak’s 2443 color
infrared film works within this photographic infrared region and is ideal
for monitoring vegetation and water.
• Electro-optical detectors on Landsat Thematic Mapper 4 and 5 are
sensitive to the reflective infrared windows from 1.55 - 1.75 m (TM
band 5) and 2.08 - 2.35 m (TM band 7).
Thermal Infrared Detectors
• Electronic detectors can also be made sensitive to photons of thermal
infrared radiant energy exiting the terrain in the two primary thermal
infrared windows: 3 - 5 m and 8 - 14 m. Sub-orbital thermal infrared
remote sensing systems utilize these spectral bands.
• The Earth’s ozone (O3) layer absorbs much of the thermal energy
exiting the terrain in an absorption band from approximately 9 - 10 m.
Therefore, satellite thermal infrared remote sensing systems usually
only record data in the region from 10.5 - 12.5 m to avoid the
absorption band.
Thermal Radiation Laws
• A blackbody is a theoretical construct that absorbs all the radiant
energy striking it and radiates energy at the maximum possible rate per
unit area at each wavelength for any given temperature.
• No objects in nature are true blackbodies, however, we may think of
the Sun as approximating a 6,000 K blackbody and the Earth as a 300 K
blackbody. If we pointed a sensor at a blackbody we would be able to
record quantitative information about the total amount of radiant energy
in specific wavelengths exiting the object and the dominant wavelength
of the object. In order to do this, we utilize two important physical
laws: the Stefan-Boltzmann law and Wien’s displacement law.
Stephen Boltzmann Law
The total spectral radiant flux exitance (Mb) measured in watts/m2 leaving
a blackbody is proportional to the fourth power of its temperature (T). This
is the Stefan-Boltzmann law and is expressed as:
M b  sT
4
where s is the Stefan-Boltzmann constant equaling 5.6697 x 10-8 W m-2 K4 m K, and T is temperature in degrees Kelvin. The total radiant exitance
is the integration of all the area under the blackbody radiation curve.
The Sun produces more spectral radiant exitance (Mb) at 6,000 K than the
Earth at 300 K. As the temperature increases, the total amount of radiant
energy measured in watts per m2 (the area under the curve) increases and
the radiant energy peak shifts to shorter wavelengths.
Blackbody Radiation
Curves for Several
Objects including the
Sun and Earth
Wein’s Displacement Law
The relationship between the true temperature of a blackbody (T) in
degrees Kelvin and its peak spectral exitance or dominant wavelength
(max) is described by Wien’s displacement law:
max
k

T
2898 m K

T
where k is a constant equaling 2898 m ˚K.
Wien’s Displacement Law
For example, the average temperature of the Earth is 300 K
(80 ˚F). We compute the Earth’s dominant wavelength as:
max
k

T
2898 m K

 9.67 m
300 K
Wien’s Displacement Law
The average temperature of an 800 K red hot object is:
max
k

T
2898 m K

 3.62m
800 K
Wien’s Displacement Law
• The dominant wavelength provides valuable information about which
part of the thermal spectrum we might want to sense in. For example, if
we are looking for 800 K forest fires that have a dominant wavelength
of approximately 3.62 m then the most appropriate remote sensing
system might be a 3-5 m thermal infrared detector.
• If we are interested in soil, water, and rock with ambient
temperatures on the earth’s surface of 300 K and a dominant
wavelength of 9.67 m, then a thermal infrared detector operating in
the 8 - 14 m region might be most appropriate.
Emissivity
• The world is not composed of radiating blackbodies. Rather it is
composed of selectively radiating bodies such as rocks, soil, and water
that emit only a fraction of the energy emitted from a blackbody at the
same temperature. Emissivity, , is the ratio between the radiant flux
exiting a real-world selective radiating body (Mr) and a blackbody at
the same temperature (Mb):
Mr

Mb
Emissivity
• All selectively radiating bodies have emissivities ranging from 0 to <1
that fluctuate depending upon the wavelengths of energy being
considered. A graybody outputs a constant emissivity that is less than
one at all wavelengths.
• Some materials like distilled water have emissivities close to one
(0.99) over the wavelength interval from 8 - 14 m. Others such as
polished aluminum (0.08) and stainless steel (0.16) have very low
emissivities.
Spectral emissivity of a
blackbody, a graybody,
and a hypothetical
selective radiator
Spectral radiant exitance
distribution of the
blackbody, graybody,
and hypothetical
selective radiator
Radiant energy exiting Water, Granite, and Dunite heated to
350 K compared with a blackbody at the same temperature
Emissivity
Two rocks lying next to one another on the ground could have the same
true kinetic temperature but have different apparent temperatures when
sensed by a thermal radiometer simply because their emissivities are
different. The emissivity of an object may be influenced by a number
factors, including:
• color -- darker colored objects are usually better absorbers and
emitters (i.e. they have a higher emissivity) than lighter colored objects
which tend to reflect more of the incident energy.
• surface roughness -- the greater the surface roughness of an object
relative to the size of the incident wavelength, the greater the surface
area of the object and potential for absorption and re-emission of
energy.
Emissivity
• moisture content -- the more moisture an object contains, the greater its
ability to absorb energy and become a good emitter. Wet soil particles
have a high emissivity similar to water.
• compaction -- the degree of soil compaction can effect emissivity.
• field-of-view -- the emissivity of a single leaf measured with a very high
resolution thermal radiometer will have a different emissivity than an
entire tree crown viewed using a coarse spatial resolution radiometer.
• wavelength -- the emissivity of an object is generally considered to be
wavelength dependent. For example, while the emissivity of an object is
often considered to be constant throughout the 8 - 14 m region, its
emissivity in the 3 -5 m region may be different.
Emissivity
• viewing angle - the emissivity of an object can vary with sensor
viewing angle.
We must take into account an object’s emissivity when we use our
remote radiant temperature measurement to measure the object’s true
kinetic temperature. This is done by applying Kirchoff’s radiation law.
Kirchoff’s Radiation Law
• Remember that the terrain intercepts incident (incoming) radiant flux
(i). This incident energy interacts with terrain materials. The amount
of radiant flux reflected from the surface (r), the amount of radiant
flux absorbed by the surface (a), and the amount of radiant flux
transmitted through the surface (t) can be carefully measured as we
apply the principle of conservation of energy and attempt to keep track
of what happens to all the incident energy. The general equation for the
interaction of spectral () radiant flux with the terrain is:
1   i   r  a  t
Kirchoff’s Radiation Law
• Dividing each of the variables by the original incident radiant flux:
 i  r a t



 i   i   i   i
allows us to rewrite the initial equation as:
1  r  a   t 
where r is spectral hemispherical reflectance by the terrain, a is
spectral hemispherical absorptance, and t is spectral hemispherical
transmittance.
Kirchoff’s Radiation Law
• The Russian physicist Kirchoff found that in the infrared portion of
the spectrum the spectral emissivity of an object generally equals its
spectral absorptance, i.e. a ~ . This is often phrased as:
“good absorbers are good emitters and
good reflectors are poor emitters”.
Also, most real-world materials are usually opaque to thermal radiation
meaning that no radiant flux exits from the other side of the terrain
element. Therefore, we may assume transmittance, t = 0. Substituting
emissivity for absorptance and removing transmittance from the
equation yields:
1  r   
Kirchoff’s Radiation Law
• This simple relationship describes why objects appear as they do on
thermal infrared imagery. Because the terrain does not lose any
incident energy to transmittance, all of the energy leaving the object
must be accounted for by the inverse relationship between reflectance
(r) and emissivity (). If reflectivity increases then emissivity must
decrease. If emissivity increases then reflectivity must decrease. For
example, water absorbs almost all incident energy and reflects very
little. Therefore, water is a very good emitter and has a high emissivity
close to 1. Conversely, a sheet metal roof reflects most of the incident
energy, absorbs very little, yielding an emissivity much less than 1.
Therefore, metal objects such as cars, aircraft, and metal roofs almost
always look very cold (dark) on thermal infrared imagery – they are
poor emitters.
Kirchoff’s Radiation Law
• The goal of thermal infrared remote sensing is to be able to point a
radiometer at an object and have the apparent radiant temperature
recorded (Trad) equal the true kinetic temperature of the object (Tkin).
Unfortunately, the radiant flux from a real-world object at a given
temperature is not the same as the radiant flux from a blackbody at the
same temperature largely due to the effects of emissivity. Knowing the
emissivity characteristics of an object makes it possible to modify the
Stefan-Boltzmann law (originally applicable to blackbodies) so that it
pertains to the total spectral radiant flux of real-world materials (Mr):
M r    s  Tkin
4
It takes into account the temperature of the object and its emissivity to
create a more accurate estimate of the radiant flux exiting an object.
Kirchoff’s Radiation Law
• Thermal infrared remote sensing systems generally record the
apparent radiant temperature, Trad of the terrain rather than the true
kinetic temperature, Tkin. If we assume that the incorporation of
emissivity in the previous equation has improved our measurement to
the point that:
Mr = s Tkin 4 and we assume that
Mb = s Trad4
and
Mr = Mb
then,
s Trad4 =  s Tkin 4
Therefore, the radiant temperature of an object recorded by a remote
sensor is related to its true kinetic temperature and emissivity by the
following relationship: Trad = 1/4Tkin
Thermal Properties of Terrain
Water, rocks, soil, vegetation, the atmosphere, and human
tissue all have the ability to conduct heat directly through
them (thermal conductivity) on to another surface and to
store heat (thermal capacity). Some materials respond to
changes in temperature more rapidly or slowly than others
(thermal inertia).
Thermal Properties of Terrain
• Thermal conductivity (K) is the rate that heat will pass through a
material and is measured as the number of calories that will pass
through a 1-cm cube of material in 1 second when two opposite faces
are maintained at 1 ˚C difference in temperature (cal cm-1 sec-1 ˚C).
The conductivity of a material is variable due to soil moisture and
particle size. Many rocks and soils are extremely poor conductors of
heat.
• Thermal capacity (c) is the ability of a material to store heat. It is
measured as the number of calories required to raise a gram of material
(e.g. water) 1 ˚C (cal g-1 ˚C-1). Water has the highest thermal capacity
(1.00). It stores heat very well relative to all the other materials.
Thermal Inertia
• Thermal inertia (P) is a measurement of the thermal response of a
material to temperature changes and is measured in calories per square
centimeter per second square root per degree Celsius (cal cm-2 sec -1/2
˚C-1). Thermal inertia is computed using the equation:
P  K  pc
where K is thermal conductivity, p is density (g cm-3), and c is thermal
capacity. Density is the most important property in this equation
because thermal inertia generally increases linearly with increasing
material density.
Apparent Thermal Inertia
• It would be wonderful if we could remotely sense each of the aforementioned
variables and then simply compute thermal inertia. Unfortunately, this is not the case
because conductivity, density, and thermal capacity must all be measured in situ.
Nevertheless, it is possible to remotely sense and compute an apparent thermal inertia
measurement per pixel in the following manner. A thermal infrared image is acquired
over the identical terrain in the nighttime and in the early afternoon. The two images are
geometrically and radiometrically registered to one another and the change in
temperature, ∆T for a specific pixel is determined by subtracting the nighttime apparent
temperature from the daytime apparent temperature. The apparent thermal inertia (ATI)
per pixel is:
1 A
ATI 
T
with A being the albedo (reflectance) measured in a visible band of the spectrum for the
pixel of interest.
Diurnal Temperature Cycle of Typical Materials
• The diurnal cycle encompasses 24 hours. Beginning at sunrise, the earth begins
intercepting mainly short wavelength energy (0.4 - 0.7 m) from the Sun. From about
6:00 am to 8:00 pm, the terrain intercepts the incoming short wavelength energy and
reflects much of it back into the atmosphere where we can use optical remote sensors to
measure the reflected energy. However, some of the incident short wavelength energy is
absorbed by the terrain and then re-radiated back into the atmosphere as thermal
infrared long wavelength radiation (3 - 14 m). The outgoing longwave radiation
reaches its highest value during the day when the surface temperature is highest. This
peak usually lags two to four hours after the midday peak of incoming shortwave
radiation, owing to the time taken to heat the soil. The contribution of reflected short
wavelength energy and emitted long wavelength energy causes an energy surplus to
take place during the day. Both incoming and outgoing shortwave radiation become
zero after sunset (except for light from the moon and stars), but outgoing longwave
radiation continues all night.
Peak Period of Daily
Outgoing Longwave
Radiation and the Diurnal
Radiant Temperature of
Soils and Rocks,
Vegetation, Water, Moist
Soil and Metal Objects
Diurnal Temperature Cycle of Typical Materials
• If all the curves for soils and rocks, water, vegetation, moist soil, and metal objects lie
exactly on top of one another, then remote sensing in the thermal infrared portion of the
spectrum would be of no value because all the phenomena would have the same
apparent radiant temperature. There would be no contrast in the imagery between the
different phenomena. Fortunately, there are only two times during the day (after sunrise
and near sunset) when some materials like soils and rocks and water have exactly the
same radiant temperature. During this crossover time period it is not wise to acquire
thermal infrared remotely sensed data.
• Fortunately, some materials store heat more efficiently that others, i.e. they have a
higher thermal capacity. For example, water has a much higher thermal capacity than
soils and rocks). Its diurnal temperature range fluctuates very little when compared with
the dramatic temperature fluctuation of soils and rocks during a 24-hr period.
Thermal Infrared Detectors
Thermal infrared detectors are usually composed of:
• In:Sb (indium antimonide) with a peak sensitivity near 5µm;
• Gd:Hg (mercury-doped germanium) with a peak sensitivity near
10 µm, or
• Hg:Cd:Te (mercury-cadmium-telluride) sensitive over the range
from 8 - 14 µm.
The detectors are cooled to low temperatures (-196 ˚C; -243 ˚C; 73 ˚K)
using liquid helium or liquid nitrogen. Cooling the detectors insures that
the radiant energy (photons) recorded by the detectors comes from the
terrain and not from the ambient temperature of objects within the
scanner itself.
Thermal Infrared Remote Sensing
There is an inverse relationship between having high spatial resolution
and high radiometric resolution when collecting thermal infrared data.
• The larger the radiometer instantaneous-field-of-view, , the longer the dwell time that
an individual detector can view the terrain within the IFOV during a single sweep of the
mirror. A larger IFOV provides good radiometric resolution which is the ability to
discriminate between very small differences in radiant energy exiting the terrain
element. In fact, the radiant energy signal measured may well be much stronger than
any noise introduced from the sensor system components. When this takes place we say
that we have a good signal to noise ratio. Of course, the larger the IFOV, the poorer the
ability to resolve fine spatial detail. Selecting a smaller IFOV will increase the spatial
resolution. But, the sensor will dwell a shorter time on each terrain element during a
sweep of the mirror, resulting in poorer radiometric resolution and perhaps a poorer
signal to noise ratio.
Pre-dawn Thermal Infrared Image of Effluent Entering the Savannah River Swamp System
Savannah
River
Savannah
River
2x reduction
March 31, 1981
4:28 am; 3 x 3 m
Pre-dawn Thermal Infrared Image of a
Residential Subdivision in Forth Worth, Texas
h
e
f
g
a
b
d
c
250 m AGL
1 mrad IFOV
6:45 am
Jan 10, 1980
0.25 x 0.25 m
Daytime Optical
and Nighttime
Thermal Infrared
Imagery of New
York City
Aerial Photograph
Thermal Infrared
Daytime Optical and
Nighttime Thermal Infrared
Imagery of the University of
South Carolina Campus
2x reduction
April 26, 1981
4:56 am
1x1m
Vertical Aerial Photograph
Solomon Blatt Fieldhouse on
the University of South
Carolina Campus
metal vent
a.
Pre-dawn Thermal Infrared Image
overhanging
eaves
b.
March 10, 1983
4:30 am
0.5 x 0.5 m
Forward Looking
Infrared (FLIR)
Examples