nthetic and
efined.
e a part of
)5/, where
pairs and
te system,
ibration of
er Vision,
lose-range
k Bundle
112), pp.
bration of
pport, 94-
imeters to
projective
ind, 12(1),
n without
nd, 13(2),
ges using
eports of
-1, Espoo,
ion Using
tical 3-D
in, Karls-
sation of
netria 37,
'ammetry,
netry and
of Metric
Ielbourne,
The Leica system for orientation of linear array sensor imagery
Azubuike G. Nwosu and Alfons Meid
Leica AG
Photogrammetry and Metrology
CH-5035 Unterentfelden
Switzerland
Phone: ++41 62 737 6821 ++41 62 737 6822 Fax: ++41 62 737 6830
email: nwag@pmu.leica.ch and meal@pmu.leica.ch
ABSTRACT
The use of linear array satellite imagery, expecially SPOT, is extensive and growing in popularity because of its stable orbit, image-quality,
continuous global coverage, and suitability for a lot of mapping tasks. This model is based on the approach of Westin where a precise orbit is
derived from the ephemeris and then used to derive a simplified orbit based on fewer parameters, so chosen that there is simpler derivation of the
dynamic parameters of orientation. Control points are used to make the required small adjustments to these parameters via bundle adjustment, with
the option of utilising tie points.
The modelling approach benefits the achievement of a precise real-time program for the Leica Mapping Terminal (LMT). This RTP has taken the
approach articulated by Kratky, and making extensive use of polynomials linking orientation and desired plate position, with modifications. The
concept of the Leica Photogrammetric Workstation, which is realised on SD2000 and SD3000 instruments as well as on upgraded DSR and Wild
AC/BC instruments, is completely unchanged by the implementation of the SPOT model.
Using seven well spaced control points, this model consistently showed RMS. errors in the 1 pixel region at check points. With more control points
the combined RMS. errors at GCPs and check points stabilised just below the pixel region. The Westin model as originally articulated is not
adequate for modelling the SPOT dynamic system. À linear parameter in phi showed the highest sensitivity, removing the misfits to the points field.
KEYWORDS: SPOT, Linear Arrays, Real-time realisation, Analytical, Block Adjustment
1.0 INTRODUCTION
The utilisation of satellite imagery, especially SPOT, has been
increasing rapidly in recent years. SPOT imagery meets all
specifications for 1:100,000 mapping and has been shown to have 80%
of the information requirements for 1:50,000 mapping (Gugan and
Dowman 1988). Space-borne data sources for mapping is expected to
increase with the imminent launching of the first orbiting platform of
the earth observation system (NASA 1993) and of a new SPOT satellite
with sensors offering better performance. This paper discusses the the
SPOT module of Leica’s linear array geometric processing system.
Leica has been involved in the development of an orientation system
for SPOT imagery for more than 10 years. Leica's SPOT MS software
offers state-of-the-art SPOT modelling in a user-friendly, graphics
driven Windows environment. It is designed with a similar interface as
other LEICA orientation software and is supported by a rigorously
computed real-time realisation for the Leica Mapping Terminal (LMT).
1.1 The SPOT Imaging System
SPOTI is a sun-synchronous, near circular orbiting satellite launched
by France in 1986,. This polar orbiting satellite carries two CCD-
imaging devices (HRVI and HRV2, High Resolution Visible) which
operate in Panchromatic (P) and multispectral (XS) modes. The two
imaging devices can operate at the same time, and each device can
operate in either of the two modes but not both modes simultaneously.
Each HRV instrument has 3 panchromatic CCD-sensors with 6000
imaging detectors spaced at 13 um. These CCD-imaging devices are
linear arrays operated in pushbroom mode. In panchromatic (PAN)
mode, readings from these sensors are integrated into one set of 6000
pixels per line; In multispectral (XS) mode, readings of succesive pairs
of detectors are added to produce 3 sets of 3000 pixels per line.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996
Measuring radiation reflected from imaged surfaces at intervals of
1.5004 micro-seconds (us) in PAN imaging results in a average of 10
metres in ground sample distance (GSD) per pixel along track. The XS
mode, samples are at at 3.008 micro-seconds with an equivalent GSD
of 20 metres. A mirror attached to the imaging devices allow for
viewing angles of up to +27 degrees, at which the GSD in
panchromatic mode could reach 13.5 metres. A stereo made of two
views separated by 45 degrees view-angle would typically give a base-
height ratio of 1. The panchromatic band covers 0.51 to 0.73 (um) in
wavelength, the multispectral bands are 0.5 to 0.59 uum (green), 0.61 to
0.68 um (red), and 0.79 to 0.89 pum (infra-red).
Spot scenes are segmented and delivered in an average GSD of 10m for
the panchromatic image or 20m for the multispectral image in 60 km x
60 km scenes. For photogrammetric processing, the required SPOT
product is processed to level 1A; only radiometric correction has been
done. Level 1AP refers to hard copy negatives with specially marked
image corners for faster and more accurate inner orientation
measurements.
2.0 THE SPOT GEOMETRIC MODEL
Many SPOT models have been developed in recent years. Many have
been based on determinated orbital parameters from ephemeris data
(Gugan and Dowman 1988, Westin 1990, Radhadevi 1994). Usually,
six independent parameters can be used to describe a Keplerian orbit;
one possible set of parameters are:
Semi-major axis (a), Eccentricity (e), Inclination (I), Right Ascension
of ascending node (2), Argument of the perigee (®) and, Mean
Anomaly (M).
The Leica model is based on Westin's approach, chosen because it is
known to be very accurate and its reduced parameter set allows for
