orbit over
pre-flight
tor sensor
id be used
ch sensor.
alibration
igned and
gets and
its during
ration site
is paper.
developed
ution IRS
OMATIC
alibration
ansfer of
nosphere.
based on
HARODI
of each
ation are
HT
WO Ways,
done by
ven often
| band. In
ne output
a uniform
Absolute
nt output
1 uniform
ration are
rd source,
in-flight
Pre-flight
ted. light
own and
rcises are
transfer
1 spectral
| are used
Pre-flight
| with on-
board calibrators so far. The problem with on-board source is
the stabilization and degradation of intensity output over a
period of time. These sources could only be used to monitor the
performance of CCD and associated electronics but the
degradation on the optics used in the payload could not be
known. Hence complete performance of the sensor could not be
evaluated. In-flight absolute radiometric calibration using
ground calibration sites offers a very good opportunity to
characterize the complete sensor in the way it collects the data.
It is for this reason this type of approach is becoming very
popular among CEOS working groups and is being used by
various workers in this field. This approach provides an end-to-
end calibration which if followed accurately, can be used to
study and monitor sensor performance over its complete life
cycle and generate new calibration coefficients. Slater P.N.
et.al.(1987) have reported reflectance based methods for in-
flight absolute calibration of multi- spectral sensors on-board
CZCS and Landsat-TM satellites. The techniques were used on
White sands, New Mexico as calibration site including
synchronous data collection by aircraft over the site. SPOT
calibration on the French test site was carried out by Santer R.
et.al.(1992). Moran M.S. et.al (1995) has retrieved reflectance
from Landsat-TM and SPOT HRV data for bright and dark
targets. Thome K.J. et.al(2000) have presented early ground
reference calibration results for Landsat-7 ETM+ using small
test sites. A generalized approach to the vicarious calibration of
multiple earth observation sensors using hyper spectral data
based on QUASAR monitoring has been described by Teillet
P.M. et.al (2001). As a part of the present work done earlier,
absolute radiometric calibration of IRS-1B and Landsat-5/TM
sensor was carried out using ground targets by Shukla A.K.
et.al(1994). Effect of physical properties of atmospheric
aerosols on path radiance for the correction of satellite counts
were studied by Nair P.R. et.al(1997).
3. PRESENT APPROACH FOR VICARIOUS
CALIBRATION
In this approach, reflectance of the target is measured along
with other atmospheric parameters such as total optical depth,
aerosol optical depth, ozone and water content during satellite
pass. The basic sun-target-sensor geometry is considered and
target radiance at the top of atmosphere is computed by
considering all aspects in the propagation of radiation through
the intervening atmosphere and reaching sensor altitude.
Essentially, scattering and absorption of radiation by the
atmosphere is accounted by the use of in-house designed
algorithm developed especially for the correction of spectrally
varying radiation. This TOA target radiance is compared with
satellite measured radiance for the same target. As mentioned
earlier, satellite measured radiance is computed using preflight
calibration coefficient based on LTC curve. The three main
components of the present methodology can be broadly
categorized as: '
e An accurate atmospheric correction algorithm
* A controlled calibration site with targets of known
reflectance
* Operationalization of well calibrated measuring
instruments
Using this approach, it is possible to carry out in-orbit
calibration and evaluation of IRS sensors routinely on an
operational basis. One of the main and important factor in
vicarious calibration is availability of easy and frequent access
to the calibration site for synchronous data collection during
IAPRS & SIS, Vol.34, Part 7, “Resource and Environmental Monitoring”, Hyderabad, India,2002
181
satellite pass. This could be realized on CHHARODI
calibration site which is 30 km. away from SAC campus.
4. SATCOR ALGORITHM FOR ATMOSPHERIC
CORRECTION
In the present approach, scattering and absorption of incoming
solar radiation and reaching satellite sensor has been accounted
based on the simplified theory of radiative transfer. These
processes modify the target reflected radiation from earth’s
surface. The atmosphere introduces an extra component in the
intrinsic radiance of target. Background reflectance plays the
major role in modifying signal from the target. The incoming
solar radiation is also modified by the atmosphere and diffuse
component is added to the direct incident radiation. The basic
assumptions followed in the transfer of radiation are :
e The atmosphere is plane, parallel and horizontally
homogeneous.
e There is no diffuse radiation entering the sky from above.
e The surface background is Lambertian.
e There is no absorption within the region where scattering
takes place.
e No clouds are present.
Two types of scattering occurs in the atmosphere namely,
Rayleigh and aerosol due to interaction of e.m. radiation with
atmospheric constituents(Mc Cartney E.J.,1976). Rayleigh
scattering takes place when the radius of scatterer is far smaller
than the wavelength of interaction and varies inversely as
fourth power of wavelength. This type of scattering gives equal
amount of radiation in forward and backward direction. When
the size of scatterer is larger than the wavelength, Mie
scattering occurs and in such a case, angular distribution of
scattered radiation is asymmetric with forward scattering
dominating. Atmospheric correction methodologies have been
dealt extensively by various workers in the past for application
in the remote sensing through satellite sensors and have been
based on the radiative transfer theory developed by
Chandrasekhar S.(1970). Simplified solutions have been arrived
by Turner and Spencer(1972).
For a nadir viewing sensor, total radiation reaching at sensor
can be expressed as :
Laisse, (1)
Where, L, is the radiance reaching satellite sensor from the
target alone and L, is path radiance. These are spectrally
dependent radiation affected by the spectral bands in which
sensor data is collected. For a Lambertian type of surface,
intrinsic reflectance R of the target is given by:
R = (7L)/Te E; (2)
In the above expression, Tg is the atmospheric transmittance in
the direction of sensor and E, is the total down-welling
irradiance on the targets corrected for sun-earth distance. Thus,
if Lp, Te, E, and R are known accurately, combining equation
(1) and (2), radiance L, at the top of atmosphere could be
computed by the following expression :
L = {RT,E,)/%+ la (3)
