anbul 2004
he camera
> rigorous,
n software
ncludes a
Took Kit
new sensor
and sensor
the sensor
k contains
adjustment
Fic camera
he interior
atform.
OF OV-3
having as
(ite takes
ed by the
hod that 1s
calibration
s involved
s that they
of work is
n members
pproach to
je: camera
ermination
the camera
in the on-
uly 26
uly 27
uly 27
ug 22
ep 05
ep 17
ep 17
jet 03
iov 04
ie Events
V-3 into its
jon the first
control of
month, the
e was out-
f tests were
]lite.
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part Bl. Istanbul 2004
4.3 First Image: 2003 July 26
The first image was collected over the east coast of the
United States. This first image began a chain of events
related to the on-orbit tuning and calibration of the camera.
Some of the events include: image quality assessment,
camera focusing, radiometric calibration and geometric
calibration.
4.4 Initial Calibration of the Attitude Determination
System: 2003 July 27
The initial calibration of the attitude determination system
represented the first major milestone in the on-orbit
geometric calibration of the satellite. This step determined
the alignments between the star trackers and the gyros. Also
determined were the on-orbit performance characteristics of
the star trackers and the gyros. The truth data used in this
calibration step was the reference stars in the star tracker’s
catalog.
Much of this calibration step was accomplished while the
camera door was closed. This calibration also included a
special maneuver that was designed to move the vehicle in
an optimum way to provide reliable alignment estimates.
4.5 Coarse Boresight Adjustment: 2003 Aug 22
This step was used to improve the boresight alignment of
the camera. Imagery from the geometric calibration range
was used to establish a boresight error of less than 100
meters.
4.6 Final Camera Focus Adjustment: 2003 Sep 05
The initial days of operating the camera were filled with
many activities, including achieving the best focus of the
camera. There is a flat focus mirror inside the camera that can
be moved by very small amounts to improve the focus. The
mirror was moved several times until the best focus was
achieved based on a metric assessment that involved the
examination of edge sharpness in the imagery [Kohm and
Tira 2004]. The movement of the mirror causes changes in
the focal length and optical distortion of the camera,
therefore only imagery collected after the focus had been set
could be used in the initial geometric calibration of the
camera.
4.7 Orbit Determination Model Tuned and Verified: 2003
Sep 17
JPL performed the tuning and verification of the orbit
determination model [Kuang 2004]. The postfit range
residual after tuning was less than | meter (RMS). The
formal sigma of orbital positioning error is the propagated
a posteriori error estimate and was less than | meter.
Another test is also performed in computing two time
periods with an overlap in time. This showed differences
that were less than 1 meter in 3D.
The following measurement and dynamic models are used in
OV3 orbit determination:
e 27 hour orbit arc (nominal)
e BENT ionosphere model, estimating one scale factor
per arc
e 70x70 gravity model from GRACE data
e Satellite model of one solar panel with one cylinder
for drag and solar pressure
e Satellite attitude determined by the quaternion
measurement
e Atmospheric drag, estimating one scale factor per arc
e Solar radiation pressure, fixed
e Empirical once-per-revolution cross-track and along-
track forces
e Stochastic radial, cross-track and along-track forces
e White noise receiver clock bias estimated at every 5
minute
e Random-walk clock offset between receiver antennas
estimated at every 20 minute
Date P1 Data Number Number
Residual of Points of
RMS (m) Outliers
03JUL06 0.96 2300 22
03JULO07 0.98 1837 21
03JULOS 0.98 2065 243
03JUL19 0.84 1832 13
03JUL20 0.86 2149 20
031071 | 0.84 1953 23
Table 2: RMS of Range Residuals from Orbit Determination
F
Sig [
ma ^| LJ
(m) |
"|
i
|
ost
E
|
0! od ic ee
0 5 10 15
Time (hour)
Figure 3: 3D Formal Sigma from Orbit Determination
4.8 Coarse Geometric Calibration of the Camera: 2003
Sep 17
A coarse geometric calibration of the camera was performed
using the imagery from the first pass over the geometric
calibration range after the focus had been set. This coarse
calibration estimated first order effects such as camera
alignment, focal length and smile distortion. Other effects,
such as scale distortion were not estimated until more image
data was available and the calibration had been refined for
the attitude determination system.
4.9 Refined Calibration of the Attitude Determination
System: 2003 Oct 03
Additional data is brought into the calibration process for
the attitude determination system to improve the reliability
of the calibration parameters. The estimated accuracy of the
alignment between the star tracker and gyro axes is 0.5
arcsecond (1 sigma).
4.10 Initial Geometric Calibration of the Camera: 2003
Nov 04
The initial geometric calibration of the camera used data
from 13 panchromatic and 2 multispectral images collected
during the period from: 2003 Sep 10 to 2003 Oct 27. Each
of the multispectral bands represented a separate image for
calibration purposes. Thus, 1348-21 images were used in
the initial geometric camera calibration. The imagery was
