48
the aerosol optical depth (AOD) and the total amount of water in a vertical path
through the atmosphere (water vapour), are unknown and need to be provided by the
user or estimated from the image for a more accurate atmospheric correction. The
sensor and surface altitude parameters are defining the length of the path through
the atmosphere that the signal being measured has taken (Fig. 5). The more atmosphere the signal passes through to the sensor, the larger the atmospheric effect
which needs to be removed from the image (Fig. 5c).
For the dynamic components of the atmosphere, specifically the AOD and water
vapour, there are various sources of information and methods that attempt to estimate those parameters from the image data itself (e.g., Masek et al. 2006). These
parameters can vary over short temporal and spatial baselines while the quality of
the atmospheric correction is highly sensitive (Fig. 6) to the correct estimation of
these parameters. They also vary as a function of the wavelength.
The AOD is correlated with visibility (in km), and the two can be transformed
from one another using the following relationship,
AOD
vis
=
+
3 9449 0 08498
.
.
There are three main sources of AOD for parameterisation of the atmospheric
model; (a) ground measurements, (b) estimates from a third party satellite or (c)
estimates from the image being corrected. As the AOD varies over short temporal
and spatial baselines (Wilson et al. 2014), estimates from the image being corrected
will be the most reliable, both spatially and temporally, and hence this is the preTable 2 List of software packages for applying an atmospheric correction using a modeled
atmosphere
Software
RT Model
Sensors
License
ATCOR-4
a
(airborne)
MODTRAN
Many – see website
Commercial
ATCOR-3
b
(satellite)
MODTRAN
Many – see website
Commercial
FLAASH
c
MODTRAN
Many – see website
Commercial
LEDAPS
d
6S
Landsat (TM, ETM+)
Free but closed source
SEN2COR
e
MODTRAN
Sentinel-2
Free but closed source
ARCSI
f
6S
Landsat (MSS, TM, ETM+,
OLI), Rapideye, SPOT5, SPOT6,
SPOT7, WorldView-2,
WorldView-3, Pleiades,
Sentinel-2
Free and Open Source
a
www.rese-apps.com/software/atcor-4-airborn
b
www.rese-apps.com/software/atcor-3-satellites
c
www.harrisgeospatial.com/docs/FLAASH.html
d
ledaps.nascom.nasa.gov
e
step.esa.int/main/third-party-plugins-2/sen2cor/
f
www.rsgislib.org/arcsi
P. Bunting
the aerosol optical depth (AOD) and the total amount of water in a vertical path
through the atmosphere (water vapour), are unknown and need to be provided by the
user or estimated from the image for a more accurate atmospheric correction. The
sensor and surface altitude parameters are defining the length of the path through
the atmosphere that the signal being measured has taken (Fig. 5). The more atmosphere the signal passes through to the sensor, the larger the atmospheric effect
which needs to be removed from the image (Fig. 5c).
For the dynamic components of the atmosphere, specifically the AOD and water
vapour, there are various sources of information and methods that attempt to estimate those parameters from the image data itself (e.g., Masek et al. 2006). These
parameters can vary over short temporal and spatial baselines while the quality of
the atmospheric correction is highly sensitive (Fig. 6) to the correct estimation of
these parameters. They also vary as a function of the wavelength.
The AOD is correlated with visibility (in km), and the two can be transformed
from one another using the following relationship,
AOD
vis
=
+
3 9449 0 08498
.
.
There are three main sources of AOD for parameterisation of the atmospheric
model; (a) ground measurements, (b) estimates from a third party satellite or (c)
estimates from the image being corrected. As the AOD varies over short temporal
and spatial baselines (Wilson et al. 2014), estimates from the image being corrected
will be the most reliable, both spatially and temporally, and hence this is the preTable 2 List of software packages for applying an atmospheric correction using a modeled
atmosphere
Software
RT Model
Sensors
License
ATCOR-4
a
(airborne)
MODTRAN
Many – see website
Commercial
ATCOR-3
b
(satellite)
MODTRAN
Many – see website
Commercial
FLAASH
c
MODTRAN
Many – see website
Commercial
LEDAPS
d
6S
Landsat (TM, ETM+)
Free but closed source
SEN2COR
e
MODTRAN
Sentinel-2
Free but closed source
ARCSI
f
6S
Landsat (MSS, TM, ETM+,
OLI), Rapideye, SPOT5, SPOT6,
SPOT7, WorldView-2,
WorldView-3, Pleiades,
Sentinel-2
Free and Open Source
a
www.rese-apps.com/software/atcor-4-airborn
b
www.rese-apps.com/software/atcor-3-satellites
c
www.harrisgeospatial.com/docs/FLAASH.html
d
ledaps.nascom.nasa.gov
e
step.esa.int/main/third-party-plugins-2/sen2cor/
f
www.rsgislib.org/arcsi
P. Bunting