International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part Bl. Istanbul 2004
y Fore view
" ENS Aft view
| Stereo |
Em 5 7 4 1 20 km
600km — €.
Fig. 5 Stereo acquisition
4. HRS PARTNERS
One particularity of this instrument is its programmatic
framework. In January 1999 CNES, Spot Image and Matra
Marconi Space (MMS which is now Astrium) signed an
Agreement about the development of this instrument in
which Spot Image is funding 54% of the cost of the
instrument. This private investment on an instrument was
quite new in the Spot program, previously paid by French,
Swedish and Belgian public funds only (excepted the SPOT 5
ground segment paid by Spot Image). In April 2002 another
Agreement was signed between CNES, the French Ministry
of Defense and Spot Image about the use and exploitation of
the HRS instrument, split between French Defense and
civilian / commercial market. Spot Image and IGN (Institut
Géographique National, the French Mapping Agency) have
also defined together a new 3D Database, named
Reference3D, mainly extracted from HRS data. (Airault,
2003; Bernard, 2004)
These agreements are reflecting the fact that the main HRS
objective is to produce DEM on (very) large areas to satisfy
dual (military and civilian) use. Therefore no original image
data will be offered on the market.
S. HRS ASSESSMENT
After a two-month in-flight commissioning phase, the use of
the satellite was given in the hands of its commercial
operator, Spot Image, on the 12" of July 2003.
All on-board equipments have been checked and declared
operational. About twenty well-equipped test areas all over
the world have been used for either radiometric or geometric
calibration. (Breton, 2002)
For the HRS payload the two telescopes, HRSI and HRS2,
were calibrated with a remaining error of less than 0.05 pixel
(Bouillon 2003). HRS absolute location accuracy, decreased
from an initial 63 m RMS value right after the
commissioning phase (July 2002), down to about 20 m RMS
(Bouillon, 2004). This result is even increased (16-m @
90%) using spatiotriangulation (bundle block adjustment on a
large area) (Airault, 2003)
First assessments of DEM accuracy derived from HRS give
good results, even better than expected: in flat or low-relief
landscape the altimetric accuracy is better than 5m RMS,
with more than 9596 good correlation. (Rudowski, 2003;
Nonin, 2003; Valorge, 2003).
It can be noticed that this HRS assessment has been
performed on very large areas (about 600 km by 200 km over
the Alps) and with high quality reference data (DEM derived
from aerial photogrammetry on Manosque area, one of the
HRS-SAP test site)
6. BACKGROUND OF THE HRS-SAP
As mentioned in the introduction the HRS-SAP is a new
initiative. Its originality is due to new status and capabilities
of the HRS instrument.
Previously, for SPOT 1 to SPOT 4, and this is also true for
the main payload of SPOT 5 (the two HRG instruments) all
images (at several geometrical processing levels) are
available to any user, through the Spot Image distribution
network.
This is not usually the case for the HRS instrument which
images are archived and processed to produce and market
Digital Elevation Models, as said previously.
Nevertheless, many scientific users asked to get HRS images
not only for deriving HRS DEMs, but also for other
photogrammetric purposes or for thematic evaluations.
It has been recognized that, beyond the operational use of
HRS, for which stereo pairs are intermediate products, some
selected scientific and/or assessment projects could be given
the permission to use original HRS images. The Earth
Science and Applications unit at CNES then proposed to set
up an international team in association with ISPRS especially
for the photogrammetric assessment of the HRS instrument.
For other scientific projects, other solutions, on a case by
case basis, are also possible.
CNES reached an agreement with ISPRS for promoting and
organizing such assessment in July 2002, agreeing that the
announcement of this common initiative should be done
during the Commission I Symposium, in Denver in
November 2002.
Manfred Schroeder, as Chairman of WG 1-2 of ISPRS
(Sensor calibration and testing), agreed to represent ISPRS
and co-chair this "HRS Study Team". The initiative was
announced during the Denver Symposium and also published
on the ISPRS website and E-mailed to about 200 potential
participants.
7. THE HRS STUDY TEAM
The HRS Study Team is a organized as follows:
A Secretariat, composed of the authors of this paper, is in
charge of the management of the Program. It is co-chaired by
Alain Baudoin for CNES and Manfred Schroeder for ISPRS.
The other members are representing CNES (Christophe
Valorge) in charge of image quality, Spot Image (Marc
Bernard) in charge of HRS production and delivery and IGN
(Véronique Rudowski) in charge of Reference Data
assessment and of result synthesis.
The Principal Investigators (PI's) are responsible for
providing reference data on selected test areas. This
Reference Data should be more accurate than the expected
HRS accuracy (5m) and include a precise DEM (derived
from photogrammetry or laser with 1-2 m Z accuracy) and/or
Ground Control Points (with X-Y-Z accuracy better than
1m).
Co-Investigators (as well as PI's) should use HRS stereo pairs
and Reference Data to produce DEM (with one or several
Inte:
—
met
acci
with
Afte
wer
was
Sec
Eac
to t
alre
The
Aix
out
(Be
par
(Ai
sel
teri
fut
wo
Fe
cal
19
201
Th
the
sci
pr
In
rec
se
pri
Ca
tes