IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 8, NO. 4, JULY 2011 661
Tomographic Imaging and Monitoring of Buildings
With Very High Resolution SAR Data
D. Reale, G. Fornaro, Senior Member, IEEE,A.Pauciullo,X.Zhu,Student Member, IEEE,and
R. Bamler, Fellow, IEEE
Abstract—Layover is frequent in imaging and monitoring with
synthetic aperture radar (SAR) areas characterized by a high
density of scatterers with steep topography, e.g., in urban environ-
ment. Using medium-resolution SAR data tomographic techniques
has been proven to be capable of separating multiple scatterers
interfering (in layover) in the same pixel. With the advent of the
new generation of high-resolution sensors, the layover effect on
buildings becomes more evident. In this letter, we exploit the po-
tential of the 4-D imaging applied to a set of TerraSAR-X spotlight
acquisitions. Results show that the combination of high-resolution
data and advanced coherent processing techniques can lead to
impressive reconstruction and monitoring capabilities of the whole
3-D structure of buildings.
Index Terms—Differential interferometric synthetic aperture
radar (DInSAR), multidimensional SAR processing, SAR tomog-
raphy, synthetic aperture radar (SAR), TerraSAR-X (TSX), 4-D
SAR imaging.
I. INTRODUCTION
I
NTERFEROMETRIC synthetic aperture radar (SAR)
(InSAR) and differential InSAR (DInSAR), particularly
multitemporal DInSAR, have been proven to be effective for
accurate scatterer localization and monitoring of displacements
[1], [2]. The high accuracy and spatial density of the mea-
surements make these techniques cost effective compared to
classical geodetic techniques, typically used in environmental
risk monitoring.
The increase of the spatial resolution provides a tangible
improvement in the monitoring capabilities: Most of the in-
ternational space agencies have hence hugely invested in the
launch of large bandwidth spaceborne SAR systems. The hard-
ware improvement must be complemented by the development
of processing techniques that are able to extract the highest
possible information content from the data. In this sense, SAR
Manuscript received October 14, 2010; accepted November 25, 2010. Date
of publication January 23, 2011; date of current version June 24, 2011. This
work was supported in part by the European Union Integrated System for Trans-
port Infrastructures Surveillance and Monitoring by Electromagnetic Sensing
(ISTIMES) project and in part by the International Graduate School of Science
and Engineering, Technische Universitaet Muenchen, Munich, Germany.
D. Reale, G. Fornaro, and A. Pauciullo are with the Institute for the Elec-
tromagnetic Sensing of the Environment (IREA), National Research Council
(CNR), 80124 Napoli, Italy (e-mail: reale.d@irea.cnr.it; fornaro.g@irea.cnr.it;
pauciullo.a@irea.cnr.it).
X. Zhu is with the Technische Universität München, Lehrstuhl für Methodik
der Fernerkundung, 80333 Munich, Germany (e-mail: xiaoxiang.zhu@bv.
tum.de).
R. Bamler is with the Remote Sensing Technology Institute (IMF), German
Aerospace Center (DLR), 82234 Oberpfaffenhofen, Germany, and also with the
Technische Universität München, Lehrstuhl für Methodik der Fernerkundung,
80333 Munich, Germany (e-mail: richard.bamler@dlr.de).
Digital Object Identifier 10.1109/LGRS.2010.2098845
tomography, also known as multidimensional (3-D and 4-D)
imaging SAR (MDI-SAR), is recognized as a powerful tech-
nique that extends interferometry.
DInSAR and persistent scatterer interferometry (PSI) assume
the presence of only a single (dominant) scattering center in
each pixel. However, SAR images of complex scenarios are
affected by the interference between the responses of scat-
terers located at different elevations (slant heights). Standard
multipass interferometric techniques “look” for the matching
between the received signal and the “multipass signature” of
a scatterer: The interference of responses may hence lead to
misdetection of persistent scatterers and to height, velocity, and
time-series measurement inaccuracies.
The layover effect causes interference between the responses
of different scatterers. Layover is particularly critical in urban
areas which are characterized by a high density of scatterers
distributed on vertical structures.
As briefly explained next, MDI-SAR allows the overcoming
of the single scatterer assumption and has opened a new sce-
nario in the 3-D target reconstruction and monitoring with SAR
systems [3], [4]. On medium-resolution systems, MDI-SAR
imaging has already been proven to be effective in separating
and monitoring scatterers in layover [5], [6].
The new generation of high-resolution SAR sensors, such
as TerraSAR-X (TSX) and the COSMO-SkyMed constellation,
allows the systematic acquisition of data with spatial resolution
reaching metric/submetric values. The preliminary analysis of
these images in dense urban areas has indicated that the reso-
lution improvement brings layover of vertical structures to be
more pronounced. On high-resolution SAR data, the interfer-
ence between scatterers on the ground and on buildings is more
frequent, and it is distributed on more pixels than on data ac-
quired by medium-resolution satellites (e.g., European Remote
Sensing (ERS) satellite or ENVISAT): The tomographic ap-
proach is a tool that allows mitigating this problem [7]. More-
over, the higher the resolution, the higher are the expectations
for 3-D reconstruction on vertical structures.
In this letter, we investigate the application of SAR tomogra-
phy to a real data set of TSX spotlight images over the city of
Las Vegas, NV. The characteristics of this data set allow clear
demonstration of the potential and the advantages offered by
the SAR tomography technique.
II. L
AYOV ER A ND TOMOGRAPHY
The imaging mechanism of radar is measuring the distances
(range) of the scatterers from the sensor. If two scatterers are
1545-598X/$26.00 © 2011 IEEE
662 IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 8, NO. 4, JULY 2011
Fig. 1. Temporal multilook image of the spotlight TSX data set over the city
of Las Vegas. The layover induced by the buildings is well recognizable.
located at different positions but share the same range, they are
imaged into the same pixel: This effect is known as layover.
In the presence of a vertical structure, such as a building,
the radar signal is affected by layover between the ground, the
façade, and possibly the roof. To have an idea of the effects of
layover, a data set of 25 TSX spotlight images from ascending
orbits over the city of Las Vegas, NV, has been considered.
This celebrated city, located over a flat area, includes a high
number of tall structures. The multitemporal averaged ampli-
tude image is shown in Fig. 1: Tall buildings are well visible,
although “folded” onto the ground toward the sensor; hence,
their responses interfere with those of the targets located below
the structures.
The tomography principle is simple: By using SAR data
acquired from different elevations, an antenna along the slant
height direction can be synthesized. The synthesized array
brings resolving capabilities on the backscattering distribution
along the elevation direction, orthogonal to the radar line of
sight, and hence, it leads to the possibility to separate contri-
butions coming from scatterers with different elevations and
interfering in the same pixel [5]. The tomographic technique
(3-D imaging) has been extended also to the time direction:
The differential-tomography technique (also known as 4-D, i.e.,
space-velocity imaging) allows the separation of interfering
scatterers and the measurement of their (possibly) different
velocities [3] and time series [4].
MDI-SAR exploits both amplitude and phase information
to reconstruct, for each pixel in the spatial (i.e., azimuth/
range) domain, the backscattering distribution along the slant
height/mean deformation velocity plane. This fact already al-
lows the improvement of performance in terms of dominant
persistent scatterer detection with respect to classical PSI that
uses only phase information [8]. In this letter, we limited the to-
mographic analysis to single (dominant) and double scatterers.
Fig. 2. Distribution of the acquisitions in the spatial/temporal baseline
domain.
To search for single and double scatterers, we used the detection
approach discussed in [9] and [10] based on the generalized
likelihood ratio test. It exploits the detector for single scatterers
in [9] in a sequential way and tests the energy contribution
of the (possible) second scatterer after the cancellation of the
dominant contribution: If this test declares the absence of the
second scatterer, a second test on the presence of only one
scatterer is carried out; see [10] for more details.
III. E
XPERIMENTAL RESULTS
The TSX spotlight acquisition mode provides resolutions of
1.1 m in azimuth and 0.6 m in slant range. We applied the MDI
technique to the area of Boulevard South, also known as “The
Strip,” where many of the largest hotels, casinos, and resorts are
located. Almost all the images are acquired with the minimum
repeat cycle of 11 days, from February 2008 to April 2009:
Fig. 2 shows the baseline distribution. We note that, except for
two acquisitions, the orbital tube is rather strict: The baseline
span (B) is only approximately 207 m. This fact results in
a poor slant height resolution of about δ
s
= λr/2B
∼
=
47 m,
corresponding to a height resolution of δ
z
= δ
s
sin(ϑ)
∼
=
27 m,
where λ, r, and ϑ are the wavelength, the distance from the
scene center, and the look angle, respectively. Superresolution
SAR tomography techniques could limit the effects of this poor
resolution [7], [12]: In this letter, however, we limited our
analysis to the classical linear tomographic approach [6].
The data set was calibrated for atmospheric phase com-
ponents estimated via the low-resolution multipass DInSAR
approach in [13] before the tomographic processing.
We focused our analysis on the block of the Mirage Hotel and
Casino. It presents a tall (about 100 m) building surrounded by
a lower flat structure (entertainment attractions) about 15–20 m
over the street level.
Many features can be pointed out by comparing, in Fig. 3,
the amplitude image of the area with an orthophotograph:
REALE et al.: IMAGING AND MONITORING OF BUILDINGS WITH VERY HIGH RESOLUTION SAR DATA 663
Fig. 3. (Top) Mirage Hotel image taken from Bing maps. (Bottom) TSX
amplitude image.
1) the folding of the building toward the sensor due to the
layover (the base of the Mirage hotel is almost vertically aligned
in the Bing and TSX images); 2) the high range resolution
distributes the response of the building over a large number
of pixels; 3) the extremely high resolving capabilities of the
TSX spotlight imaging that allows distinguishing floors on the
southern façade.
A. Single-Scatterer Analysis
In Fig. 4 (top), we show the residual topography (i.e., the
topography estimated after the subtraction of the external dig-
ital elevation model—in our case, Shuttle Radar Topography
Mission) resulting from the MDI, followed by the single scat-
terer detection algorithm in [9], which tests the presence of a
persistent scatterer based on energy content along the direction
of the peak of the tomographic reconstruction. The building
rising toward the sensor is well recognizable in the detected
scatterers. As for previous analyses of TSX data [8], the density
of the detected points is also impressive.
Some considerations are now in order: First, on the southern
façade, many blue points corresponding to the ground are
detected and are mixed to scatterers colored from green to red,
corresponding to the vertical structure of the hotel. This fact
testifies that the interference in the façade and ground is very
likely. Second, in the upper right part of the image, two straight
black strips (almost aligned to the azimuth) appear clearly.
These areas correspond to two shadowing areas caused by small
Fig. 4. (Top) Residual topography and (bottom) mean deformation velocity
estimated by means of SAR tomography for the single-scatterer analysis.
Fig. 5. (Upper left image) Daily averaged temperature of the area. (Lower left
image) Residual phases after topography calibration for pixel A. (Right image)
Scatter plot.
steps (a few meters high) on the roof of the surrounding struc-
ture. One of these shadow strips falls in the radar image areas
under the layover of the north façade of the hotel. It is interest-
ing to notice in this area the presence of a high density of scat-
terers (green pixels) on the part of the façade that falls over the
shadowed strips (see the white box in Fig. 4): This high density
is the result of the absence of any interference with the ground.
The deformation map presented at the bottom of Fig. 4 also
shows an interesting phenomenon: While all the rest is stable,
the roof appears moving toward the sensor at about 2 cm/year.
For one of these apparently inflating scatterer (A in Fig. 4),
the phase signal obtained after the compensation of the topo-
graphic signature is shown in the lower left image in Fig. 5: This
plot highlights the presence of a seasonal motion, and hence, the
mean velocity is only in part able to explain this movement. The
average daily temperatures of the area, provided by the Univer-
sity of Dayton database [14], are shown in the upper left image
in Fig. 5. The high degree of correlation with the deformation
is evident; see also the scatter plot in the image on the right. As
664 IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 8, NO. 4, JULY 2011
Fig. 6. (Upper row) Residual topography and (bottom row) mean deformation velocity estimated by means of the SAR tomography for (left column) single
scatterers, (middle column) lower layer of double scatterers, and (right column) the upper layer of double scatterers.
Fig. 7. Three-dimensional visualization of the single and double scatterers reconstructed with SAR tomography on Google Earth. The color is associated to the
estimated height.
REALE et al.: IMAGING AND MONITORING OF BUILDINGS WITH VERY HIGH RESOLUTION SAR DATA 665
can be seen, thermal dilation provides contributions leading to a
mismatch with the linear displacement model that is commonly
adopted in the detection of scatterers [15]. This aspect is the
subject of future investigations.
B. Double-Scatterer Analysis
The assumption, made by the classical interferometric tech-
niques, of a single scatterer per pixel neglects the interference of
scatterers. We therefore applied the detection scheme described
in [10] which is able to test the presence of single and double
scatterers. In Fig. 6, the results of this detection algorithm are
presented with the colors coded accordingly to (upper row)
the estimated topography and (bottom row) mean deformation
velocity. These figures show the capability of the tomographic
approach to “separate” the interfering layers associated with
the ground and the façade of the building. The images on the
left column show the detected single scatterers, whereas in
the middle and right columns, the images are associated with
the ground and top layer extracted from double-scatter results,
respectively. The effectiveness of tomography in scatterer pair
separation on this layover (distributed over several range pixels)
is particularly evident in the topography reconstruction; see the
homogeneity of blue color of the ground scatterer layer and the
gradation of colors on the layer corresponding to the façade.
The high density of detected double scatterers that fills the lack
of the single scatterers analysis should be noticed.
A further confirmation of the results is provided by the
shadow stripe highlighted by the white box in Fig. 4: As
expected, no double scatterers were detected in this area.
For what concerns the deformation maps shown in the right
column in Fig. 6, by analyzing both the estimated mean de-
formation velocity and the previously estimated topography, it
is interesting to note the presence of few pixels showing an
estimated velocity that is fully congruent with that of the single
scatterers affected by strong thermal dilation.
Finally, the 3-D view of the building is shown in Fig. 7
to demonstrate the impressive potential of the new (high-
resolution) sensor generation and the potential of SAR tomog-
raphy for urban area analysis. It shows a 3-D view of the Mirage
Hotel in Google Earth obtained with the identified single and
double scatterers and without the use of the optical Google 3-D
model of the building as background: The different floors are
well visible in the left façade. The results show that these SAR
sensors orbiting hundreds of kilometers from the Earth can
provide accurate 3-D reconstruction and monitoring of single
buildings.
IV. C
ONCLUSION
High-resolution SAR systems, such as TSX and Cosmo-
SkyMed, provide an obvious improvement in the imaging
capabilities. However, specific problems associated with the
geometry of SAR become more evident: Layover is among
them, and it affects particularly the images of urban areas. By
processing spotlight TSX data, in this letter, we have shown that
SAR tomography can solve this problem and allow accurate
3-D reconstruction and monitoring. Layover associated to tall
buildings and distributed over several pixels was successfully
resolved.
Whereas layover is solvable by using, as shown, SAR to-
mography, no solutions are available for shadowing. Hence,
small incidence angles are preferred for imaging urban areas
to “pierce” areas with high density of buildings and reduce
shadowing.
R
EFERENCES
[1] P. Berardino, G. Fornaro, R. Lanari, and E. Sansosti, “A new algorithm
for surface deformation monitoring based on small baseline differential
SAR interferograms,” IEEE Trans. Geosci. Remote Sens., vol. 40, no. 11,
pp. 2375–2383, Nov. 2002.
[2] A. Ferretti, C. Prati, and F. Rocca, “Nonlinear subsidence rate estima-
tion using permanent scatterers in differential SAR interferometry,” IEEE
Trans. Geosci. Remote Sens., vol. 38, no. 5, pp. 2202–2212, Sep. 2000.
[3] F. Lombardini, “Differential tomography: A new framework for SAR
interferometry,” IEEE Trans. Geosci. Remote Sens., vol. 43, no. 1, pp. 37–
44, Jan. 2005.
[4] G. Fornaro, D. Reale, and F. Serafino, “Four-dimensional SAR imaging
for height estimation and monitoring of single and double scatterers,”
IEEE Trans. Geosci. Remote Sens., vol. 47, no. 1, pp. 224–237, Jan. 2009.
[5] G. Fornaro, F. Serafino, and F. Lombardini, “3D multipass SAR focus-
ing: Experiments with long-term spaceborne data,” IEEE Trans. Geosci.
Remote Sens., vol. 43, no. 4, pp. 702–712, Apr. 2005.
[6] G. Fornaro, F. Serafino, and D. Reale, “4D SAR imaging: The case study
of Rome,” IEEE Geosci. Remote Sens. Lett., vol. 7, no. 2, pp. 236–240,
Apr. 2010.
[7] X. Zhu and R. Bamler, “Very high resolution spaceborne SAR tomogra-
phy in urban environment,” IEEE Trans. Geosci. Remote Sens., vol. 48,
no. 12, pp. 4296–4308, Dec. 2010.
[8] S. Gernhardt S, N. Adam, M. Eineder, and R. Bamler, “Potential of
very high resolution SAR for persistent scatterer interferometry in urban
areas,” Ann. GIS, vol. 16, no. 2, pp. 103–111, Jun. 2010.
[9] A. De Maio, G. Fornaro, and A. Pauciullo, “Detection of single scatterers
in multi-dimensional SAR imaging,” IEEE Trans. Geosci. Remote Sens.,
vol. 47, no. 7, pp. 2284–2297, Jul. 2009.
[10] A. De Maio, G. Fornaro, A. Pauciullo, and D. Reale, “Detection of double
scatterers in SAR tomography,” in Proc. IEEE IGARSS Conf.,CapeTown,
South Africa, Jul. 13-17, 2009.
[11] X. Zhu and R. Bamler, “Tomographic SAR inversion by L1-norm
regularization—The compressive sensing approach,” IEEE Trans. Geosci.
Remote Sens., vol. 48, no. 10, pp. 3839–3846, Oct. 2010.
[12] A. Budillon, A. Evangelista, and G. Schirinzi, “Three-dimensional SAR
focusing from multipass signals using compressive sampling,” IEEE
Trans. Geosci. Remote Sens., pp. 1–12, to be published.
[13] G. Fornaro, A. Pauciullo, and F. Serafino, “Deformation monitoring over
large areas with multipass differential SAR interferometry: A new ap-
proach based on the use of spatial differences,” Int. J. Remote Sens.,
vol. 30, no. 6, pp. 1455–1478, Apr. 2009.
[14] Average Daily Temperature Archive. [Online]. Available: http://www.
engr.udayton.edu/weather/
[15] X. Zhu and R. Bamler, “Very high resolution SAR tomography via com-
pressive sensing,” in Proc. ESA FRINGE Workshop Adv. Sci. Appl. SAR
Interferometry, Frascati, Italy, 2009.
