Journal of Global Positioning Systems (2005)
Vol. 4, No. 1-2:82-87
GPS Campaigns for Validation of InSAR Derived DEMs
InSu Lee, Hsing-Chung Chang, Linlin Ge
School of Surveying and Information Systems, University of New South Wales, Sydney, Australia
e-mail: [e-mail] Tel: + 61 2 9385 4190: Fax: [phone]
Received: 6 December 2004 / Accepted: 26 October 2005
Abstract. Interferometric Synthetic Aperture Radar
(InSAR) is a rapidly evolving technique. Spectacular
results that are obtained in various fields, such as the
monitoring of earthquakes, volcanoes, land subsidence
and glacier dynamics, as well as in the construction of
Digital Elevation Models (DEMs) of the Earth's surface
and the classification of different land types, have
demonstrated its strength.
As InSAR is a remote sensing technique, it has various
error sources due to the satellite positions and attitude,
atmosphere, and others, so it is important to validate its
accuracy, especially for the DEM derived from SAR
images before it can be used for various applications such
as disaster prevention, flood mapping, and emergency
map.
In this study, Real Time Kinematic (RTK) GPS
positioning and Kinematic GPS positioning were chosen
as tools for the validation of InSAR derived DEM. The
results showed that Kinematic GPS positioning had
greater coverage at field test, i.e. larger number of usable
sampling points than RTK GPS. However, tracking
satellites and transmitting a data between reference-rover,
under trees are still pending tasks to be overcome in GPS
positioning techniques. Additionally, Airborne Laser
Scanning (ALS) is expected to be an alternative as an
effective tool for the validation of DEMs.
Key words: Interferometric Synthetic Aperture Radar
(InSAR), Digital Elevation Model (DEM), Real Time
Kinematic (RTK) GPS positioning, Kinematic GPS
positioning
1. Introduction
A DEM measures the height of terrain above a reference
datum. DEM as a term is in widespread use and generally
refers to the creation of a regular array of elevations,
normally squares or hexagon pattern, over the terrain (El–
Sheimy, 1999).
Nowadays DEMs can be generated with several methods
such as ground surveys, photogrammetry (e.g., analytic
photogrammetry and digital photogrammetry), InSAR
technique and Airborne Laser Scanning (ALS).
The ground surveys (GPS positioning, levelling, etc)
provide height information to a high degree of accuracy,
but are time-consuming, laborious and costly, and
provide information on point basis only. The point
information on height may not be sufficient for
conducting an engineering study on regional basis that
requires dense spatial information. The spatial extent of
height can be obtained from DEM.
The photogrammetric DEMs can be stereo-compilation
methods, automatic collection of elevation data by digital
correlation from digitized film or digital imagery, and
hybrid approaches (Molander, 2004).
Recently Shuttle Radar Topography Mission (SRTM)
obtained elevation data on a near-global scale to generate
the most complete high-resolution digital topographic
database of the Earth. SRTM consisted of a specially
modified radar system that flew onboard the Space
Shuttle Endeavor during an 11-day mission in February
of 2000. This configuration produced the single-pass
interferometry and during this period, SRTM mission
imaged the Earth’s entire land surface between 60
degrees north and 50 degrees south. The C-band SRTM
data is being processed into DEMs on a continent-by-
continent basis (Peltzer, 1999)
With the advent of InSAR, it may now be possible to
obtain height information on regional basis thereby
producing DEM up to meter level accuracy. Due to this,
the technology is gaining its momentum in many
application areas such as lithospheric movements in
geology, crustal deformation studies in seismology,
global volcano monitoring, landslide monitoring, ice and
glacial studies (Arora et al, 2002).

Lee et al.: GPS Campaigns for Validation of InSAR Derived DEMs 83
The main aim of this paper is to provide the availability
overview of GPS positioning for assessment of DEMs
and to reveal the related problems.
2. Basic Concept of InSAR and GPS positioning
2.1 InSAR Overview
Synthetic Aperture Radars (SAR) produce all weather,
day and night, high resolution images of the Earth's
surface providing useful information about the physical
characteristics of the ground and of the vegetation
canopy, such as surface roughness, soil moisture, tree
height and bio-mass estimates. By combining two or
more SAR images of the same area, it is also possible to
generate elevation maps and surface change maps with
unprecedented precision and resolution. This technique is
called “SAR interferometry”. With the advent of
spaceborne radars, SAR interferometry has been applied
to the study of a number of natural processes including
earthquakes, volcanoes, glacier flow, landslides, and
ground subsidence (Peltzer, 1999).
Fig.1 presents imaging geometry for a repeat-pass
interferometer. One interferogram is formed with images
acquired from positions A1 and A2. Assume two
identical antennas, A1 and A2, are receiving radar echo
signals from a single source. The path length
difference, ρ∆ , of the signals received by the two
antennas is approximately given by
α)Bsin(θ
1
ρ
2
ρ∆ρ −≈−=
G
G
(1)
where i ρ
G
indicates the vector from antenna i to the
target, B is the length of the baseline vector which is the
vector pointing from antenna 1 to antenna 2,
θ
is the
desired elevation (or) look angle and the baseline
orientation angle, α is the angle the baseline vector
makes with respect to the horizontal. If a ground
resolution element scatters identically for each
observation, then the difference of the two phases
depends only on the path length difference. The range
difference, ρ∆ , may be obtained by measuring,
φ
, the
phase between two interferometer signals, using the
relation
λ
ρ∆π
−=φ m2 , 2,1m = (2)
where λ is the radar wavelength and
m
equals to 1 when
the path length difference is associated with one way
difference, or 2 for the two-way path difference. Using
the simplified geometry of Fig. 1, the height of a
target, t h is given by
)cos(hh t
θ
ρ
−
=
(3)
where his the altitude of the radar antenna and
ρ
is the
slant range from the antenna to the target. Generation of
accurate topographic maps using radar interferometry
places stringent requirements on the knowledge of the
platform and baseline vectors (Hensley et al., 2001).
Fig. 1 Radar Interferometric geometry
Fig. 2 Overview of Kinematic GPS positioning
2.2 GPS positioning tec h n i ques
Kinematic GPS positioning is productive in that the
greatest number of points can be determined in the least
time. In kinematic GPS positioning, the unknown rover
was positioned 'relative' to a reference station that
occupied a point of known 3-D coordinates. Fig. 2
presents the graphic presentation of Kinematic GPS
positioning.
The Kinematic technique requires the resolution of the
phase ambiguities. There are lots of ambiguity resolution
techniques for the kinematic case. One of them is called
On the Fly (OTF).This solution required an instantaneous
positioning (i.e., for a single epoch). The main problem is
to find the positions as fast and accurate as possible. This
is achieved by starting with approximations for the

84 Journal of Global Positioning Systems
positions and improving them using least squares
adjustments or search techniques (Hoff-mann et al, 1997).
RTK GPS is the dynamic GPS positioning technique
available. Using short observation times, this system
provides precise results instantaneously whenever
continuous four-satellite tracking is available. Nowadays
kinematic carrier phase-based positioning can be carried
out in real-time if an appropriate communications link is
provided over which the carrier phase data collected at a
static base receiver can be made available to the rover
receiver's onboard computer; to generate the double-
differences, resolve the ambiguities and perform the
position calculations (Rizos, 1999). This is termed as
Real Time Kinematic (RTK) GPS positioning.
3. Generati on o f InS AR DEM and GP S c ampaigns
3.1 InSAR DEM
Interferometric SAR is now established as a method for
generating DEM from complex SAR data. Validation of
such InSAR derived DEMs is still in progress and some
results are founded in literature (Balan and Mather,
1999). Interferometry is a technique that interprets the
phase difference between two identical SAR images of a
single area taken one or more repeat orbit cycles apart.
The two ERS satellites operated in tandem for a time, and
this allowed for the collection of excellent interferometric
pairs.
In this paper, the InSAR DEM was derived using the
images acquired during tandem mission of the ERS-1
(20/10/1995) and ERS-2 (30/10/1995) satellites, where
there was only one-day difference between the acquisitions of
two radar images. Fig. 3 and 4 presents the procedure for
DEM generation from both SAR Images and InSAR
derived DEM, respectively.
Fig. 3 Generation of InSAR DEM
Fig. 4 InSAR Derived DEM
3.2 GPS Field Observation
In theses GPS campaigns, a pair of Leica SR530
receivers with firmware allowing dual-frequency and
OTF technique, essential for RTK GPS, and a pair of
AT502 antennas, L1/L2 microstrip built-in ground-plane,
and a pair of radio modem for transmitting data between
a reference station and a rover were employed. This
campaign was conducted at the Mining site, Appin, in
Australia.
For RTK GPS and Kinematic GPS positioning, a
reference station was set up at a site that had a good view
to track satellites during the period of test, and a rover
Master
SAR Ima
g
e
Slave
SAR Ima
g
e
Ima
g
e Re
g
istration
Interfero
g
ra
m
Coherence Ma
p
Phase Unwra
pp
in
g
Geocodin
g
Di
g
ital Terrain Model

Lee et al.: GPS Campaigns for Validation of InSAR Derived DEMs 85
moved along the motorway of test field. Positions of a
rover antenna were recorded every 1 second in the
receiver in real-time with accuracy in several centimeters.
At the same time, raw data of both antennas were also
stored in the receiver for post processing. With these data,
Kinematic GPS positioning was processed. A reference
station and a rover set up on the roof of vehicle are shown
In Fig. 5.
Fig. 5 A reference station and a rover
4. Analysis of GPS Obser va ble and Assess men t of
DEM ac curacy
4.1 Kinematic GPS positioning and RTK GPS
First of all, the coverage of test area between Kinematic
GPS positioning and RTK GPS and was evaluated
according to the number of usable sampled points. Fig. 6
(a) and (b) indicate the display map of points acquired
from Kinematic GPS positioning and RTK GPS,
respectively and Fig. 6(c) presents the overlaid points of
Kinematic GPS positioning and RTK GPS with an aerial
photograph as background.
(a)
(b)
(c)
Fig. 6 Points of Kinematic GPS positioning (a), RTK GPS (b), and the
overlaid map of both (c)
It seems that there is no much difference of point
coverage between Kinematic GPS positioning and RTK
GPS in Fig. 6, because some measurements recorded in
receivers while a vehicle was stationary were already
excluded in statistical analysis. However, in actuality,
there is a wide difference of data coverage between these
two methods.
Especially, some areas marked as circle and square in
Fig. 6(c), showed the different data coverage between
Kinematic GPS Positioning and RTK GPS. Kinematic
GPS positioning has about two times as many usable
sampling points as RTK GPS. This may be due to the
interference of radio linkage between reference-rover,
leading no position solution (e.g. in area, marked as
square in Fig.6(c)), and the initialisation problem, leading
no solution in RTK GPS (e.g., in area, marked as circle in
Fig.6(c)). There is also some probability of both aspects
in some areas.
The RMSE of height differences between Kinematic GPS
positioning and RTK GPS is within several centimeters.
This error value might be good as is the case with both
methods.

86 Journal of Global Positioning Systems
4.2 Assessment of DEMs’ ac c uracy
In this paper, 1 arc-second photogrammetric DEM and
ERS-1/2 Tandem InSAR DEM as space-borne radar have
the pixel size of 30m and 20m, respectively, and SRTM
DEM as shuttle-borne radar has the pixel sizes of about
90m. And GPS height profiles were used as ground truth
data.
Comparison of three DEMs, i.e. 1 arc-second
photogrammetric DEM, SRTM DEM, ERS-1/2 Tandem
InSAR DEM against GPS height profiles was used. For
this, each height profile of three DEMs was extracted
along the same locations where the sampling points in
Kinematic GPS positioning were collected. And height
profiles of Kinematic GPS positioning were chosen as
ground truth data in that the Kinematic GPS positioning
had more number of usable sampling points than RTK
GPS.
Table 1 indicates the RMSE of height difference between
three DEMs and GPS height profiles according to routes.
Tab. 1 RMSE of height difference between three DEMs and GPS height profiles
Sensors
Routes Photogrammetric DEM
SRTM DEM
(Shuttle-borne Radar)
ERS-1/2
Tandem InSAR DEM
(Space-borne Radar)
R1 2.95m 1.57m 18.01m
R2 3.26m 2.02m 30.27m
R3 2.16m 3.18m 17.31m
R4 3.24m 1.85m 14.81m
DEMs Assessment - Route 1
190.00
200.00
210.00
220.00
230.00
240.00
250.00
260.00
270.00
050100 150 200 250
Epoch
Height (m)
Kinematic GPS
Ph otog rammetric DEM
Tandem InSAR DEM
SRTM DEM
(a)
DEMs Assessment - Route 2
120.00
140.00
160.00
180.00
200.00
220.00
240.00
0100 200 300 400 500 600
Epoch
Height (m )
Kinematic GPS
Ph otog ram metric DEM
Tandem InSAR DEM
SRTM DEM
(b)
DEMs Assessment - Route 3
120.000
140.000
160.000
180.000
200.000
220.000
240.000
050100 150 200 250 300 350 400 450 500
Epoch
Height (m)
Kinematic GPS
Ph otog rammetr ic DEM
Tandem InSAR DEM
SRTM DEM
(c)
DEMs Assessment - Route 4
130.00
150.00
170.00
190.00
210.00
050100 150 200 250 300 350400 450 500
Epoch
Height (m)
Kinematic GPS
Photogrammetr ic DEM
Tandem InSAR DEM
SRTM DEM
(d)
Fig. 7 Comparison of height profiles of three DEMs against Kinematic
GPS along (a) Route1, (b) Route2, (c) Route 3, and (d) Route 4
The height profiles derived from the photogrammetric
DEM and STRM DEM have the mean RMSE of about
2.90m and 2.16m, respectively, while Tandem InSAR
DEM has the mean RMSE of about 20.10m against
GPS height profiles. In case of Tandem InSAR DEM,
this value is likely to be accepted when considering the
vertical resolution of ERS images.

Lee et al.: GPS Campaigns for Validation of InSAR Derived DEMs 87
Fig. 7 shows that three DEMs have similar trend of
heights. Especially, big turbulence of height profile
between three DEMs and Tandem InSAR DEM
occurred at Route 2. This may be due to satellite
inherent errors (e.g., positions and orientations of the
satellite), phase unwrapping errors, and atmospheric
errors, etc.
Therefore, the detailed information such as satellite
orbit information, phase unwrapping algorithm, and
especially tropospheric delay to improve the accuracy
of InSAR derived DEM is required, and more
powerful method like ALS that can validate the
accuracy of InSAR derived DEMs should be
introduced.
5. Conclusi on
This paper dealt with the validation of InSAR derived
DEM against GPS height profiles as ground truth
data. The results showed that Kinematic GPS
positioning had better coverage at the field test, i.e.
larger number of usable sampling points than RTK
GPS. Therefore, it is expected that Kinematic GPS
positioning plays an important role in the validation of
InSAR derived DEM because of its cost-effectiveness.
But the interference of radio linkage between
reference-rover, the tracking satellites and multipath
error near and/or under trees are still pending problems
to be solved.
Network-Based RTK GPS and the integration SAR
with ALS will be an alternative, and what is the most
important is that researches related to validation of
DEM are further required.
Acknowledgements
This work was supported by the Post-doctoral
Fellowship Program of Korea Science & Engineering
Foundation (KOSEF).
References
Arora M.K., Patel V. (2002): SAR Interferometry for DEM
Generation, (Available in
http://www.gisdevelopment.net/technology/rs/techrs002
1pf.htm)
Balan P., Mather P.M. (1999): Evaluation of the accuracy
on InSAR DEM as a function of baseline, wavelength
and resolution, Fringe’99 Workshop, Advancing ERS
SAR Interferometry from Applications towards
operations, Liège, Belgium, 10~12, November.
Crosetto M., Aragues F.P. (1999): Radargrammetry and
SAR interferometry for DEM generation: validation
and data fusion, CEOS’99 SAR Workshop, ESA-CNES
Toulouse, 26-29, October.
Crosetto M. (2002): Calibration and Validation of SAR
interferometry for DEM generation, ISPRS Journal of
Photogrammetry & Remote Sensing, 57: 213-227
El-Sheimy N., Valeo C., Habio A., Valeo C. (2005): Digital
Terrain Modeling: Acquisition, Manipulation And
Applications, Artech House, United Kingdom, 272p.
Hensley S., Riadh M., Paul R. (2001): Interferometric
Synthetic Aperture Radar (IFSAR). In: David F.
Maune (Ed.), Digital Elevation Model Technologies and
Applications: The DEM Users Manual, ASPRS, Chapter
6, 143-200.
Hofmann-Wellenhof B., Lichtenegger H., Collins J. (1997):
GPS Theory and Practice (4th Revised Ed.), Springer,
Wien New York, 389 p.
Kimura H., Yamaguchi Y. (2000): Detection of landslide
areas using satellite radar interferometer,
Photogrammetric Engineering and Remote Sensing,
JOURNAL OF THE AMERICAN SOCIETY FOR
PHOTOGRAMMETRY AND REMOTE SENSING,
66(3): 337-344.
Lan H. (1996): Development of Real-Time Kinematic
System Design, Performance and Results, .MSc Thesis,
Calgary, Alberta, University of Calgary, pp.16.
Manue D.F., Kopp S.M., Crawford C.A., Zervas C.E. (2001):
Introduction, In: David F. Maune (Ed.), Digital
Elevation Model Technologies and Applications: The
DEM Users Manual, ASPRS, Chapter 1, p. 1-31.
Massonnet D., Feigl K.L., Rossi M., Vadon H. (1996):
Coseismic deformation field of the M 6.7 Northridge,
California Earthquake of January 17, 1994, Recorded
by Two Radar Satellites using Interferometry,
Geophysical Research Letters, Vol. 23, No. 9, 969-972.
Molander C.W. (2001): Photogrammetry. In: David F.
Maune (Ed.), Digital Elevation Model Technologies and
Applications: The DEM Users Manual, ASPRS, Chapter
5, p.121-141.
Peltzer G. (1999): Crustal Deformation Studies using SAR
Interferometry, (Available in http://www-radar.jpl.
nasa.gov/sect323/InSar4crust/home.html).
Rizos C. (1999): GPS enhancements. In: Chris Rizos (Ed),
Notes on Basics GPS Positioning and Geodetic
Concepts, School of Surveying and Spatial Information,
University of New South Wales, pp. 15~16, Electric
version (Available in
http://www.gmat.unsw.edu.au/snap/gps/gps_notes.htm).