A
list of questions we get asked most frequently are listed
below. If however you require any additional information
on the Terrafirma service or InSAR techniques
in general, please email your request to enquiries@terrafirma.eu.com.
SAR SYSTEMS
1. What
are SAR systems?
2. How do SAR systems work?
3. Why doesn’t a radar image look like a “real world”
image?
4. What are the problems affecting SAR systems?
5. What are the main SAR platforms?
6. What are the characteristics differencing the
various SAR systems?
7. What are the differences between
radar and optical sensors?
8. What is the difference between
slant range and ground range?
9. Why are SAR images called complex?
10. What is SAR data used for?
11. What are the interferometric techniques and
interferograms?
12. What are the output products
of interferometry?
13. Under which conditions can interferometry
be applied?
14. What are the limitations of interferometry?
15. What does baseline mean?
16. How can interferograms be used in geophysics?
17. What kind of geophysical deformations can be
detected in an interferogram?
MULTI INTERFEROGRAM TECHNIQUES
AND PS TECHNIQUE
18. What is the PS Technique?
19. What does the PS Technique consist in?
20. What are PSs?
21. What is the density at which PS are generally
retrieved?
22. What is the difference between the PS Technique
and traditional differential SAR interferometric techniques?
23. How are, in detail, atmospheric effects removed
from displacement measurements?
24. How and at what stage of the processing are
non linear movements taken into account?
25. What do we mean by interferogram stacking?
26. Why are sometimes multiple PS retrieved, which
are disposed along a straight line?
27. What are the processing steps which are not
performed automatically but require an expert operator?
28. Can the Technique be applied to steep relief
areas?
29. Which radar data is used as basis for processing
with the PS Technique?
30. Can data from different satellites
be used together to create differential interferograms
and apply the PS Technique?
31. What are the criteria for the
selection of processable images?
PRECISION ASSESSMENT AND PRACTICAL
PROBLEMS
32. How much accurate are the PS movements measured?
33. How is the accuracy of measurements related
to the images used?
34. How much precise is the location
of the PS points retrieved?
35. What kind of ground movements
can be measured by the PS Technique?
36. What are the limitations of the PS Technique?
37. Why is the precision of PS displacement velocities
in the order of millimetres and the precision of their
heights in the order of decimetres?
38. Can the density of PS over a new site be predicted?
39. Is there a way to increase the number of PS
retrieved over an area of interest?
40. Why should I resort to the PS Technique?
41. What are the applications of the PS Technique?
42. Does the kind of satellite data used influence
the results obtained through the PS Technique?
COMPARISON WITH STANDARD TOOLS
43. What is the difference between the PS Technique
and GPS?
44. What is the difference between the PS Technique
and optical levelling?
45. How can PS Technique results
be calibrated?
SAR SYSTEMS
1. What are SAR systems?
SAR systems are
satellite sensors working with microwaves. SAR stands
for Synthetic Aperture Radar and is so called because
it employs a small physical antenna, instead of the large
antennas used in Real Aperture Radars (SLR). The antenna
is used to collect all energy coming from the ground.
A larger antenna means improved resolution in SAR images.
SAR systems are able to synthesize a large antenna even
if they use a small one by employing the Doppler principle
and certain data-processing techniques. The synthetic
aperture is the distance covered by the platform during
the time interval between the first and the last acquisition
of a certain object on the ground. A SAR consists basically
of an emitter of microwave radiation and a receiver. All
modern satellite and air-borne radars use the SAR mode,
because it does not require carrying a bulky large antenna.
2.
How do SAR systems work?
The microwave radiation is
emitted by the sensor in short pulses with a certain angle
(look angle is the angle between the nadir and the direction
in which signals are emitted), reaches the ground with
a certain wavelength and polarization and is reflected
back by the various objects. During the reflection and
the passage through the atmosphere the microwaves change
their phase, amplitude and polarization. The signal is
then received by the antenna on the satellite platform
and registered in terms of phase, polarization and amplitude.
The record is finally processed to produce an image covering
a certain area, whose size depends on the satellite sensor.
3.
Why doesn’t a radar image look like a “real world” image?
Radar images represent in
greyscale how intensely objects on the Earth surface reflect
one particular wavelength. The wavelengths used are not
the ones we are used to see with our eyes. Furthermore,
each image represents only one wavelength acquired in
one polarization mode. True color images are instead created
as a mix of three bands when a different color (red, green
or blue) is associated to each of them. New radar platforms
are planned for the next future (the Canadian Radarsat-2
and the Japanese ALOS) and they will carry sensors able
to detect the signal backscattered from the ground in
four different modes contemporarily: with HH, VV, HV and
VH polarization. That means it will be possible to have
four images of the same area acquired at the same time
and to create false colors compositions.
Some rules exist that can
help interpreter certain features on radar images. For
instance, regions of calm and smooth water appear black
because they reflect the radar signal away from the receiver.
Human made structures like buildings, bridges, metallic
structures result in bright spots because they strongly
reflect the signal toward the receiver. Hills and slopes
are bright on the side reached by the microwave radiation
and dark on the opposite side, not reached by the radiation.
4.
What are the problems affecting SAR systems?
SAR sensors, like all radar sensors, are subject to geometric
limitations due to the side looking nature of the illumination.
Characteristics affecting most severely radar images are:
shadows, foreshortening, variation in pixel resolution
over the image, layover, speckle.
Shadows are areas on the image which give a weak return
to the recorder because they face away from the radar
signal emitter. Shadows increase at the extremes of the
image because of the wider look angle.
Foreshortening is the effect
resulting in a compression of the slopes facing perpendicularly
the emitted signal. The variation in pixel resolution
occurs because of the increase in look angle as the distance
from nadir increases; i.e. as we move far from nadir,
the same variation in degrees in look angle corresponds
to an increasing number of meters on the ground.
In case of features on the
ground extending in height, the signal emitted by the
satellite will reach first the top of the feature and
later the bottom of it. In the resulting radar image,
this will cause the vertical feature to appear as if it
was leaning toward the nadir. For this reason the effect
is called layover.
Speckle is an effect causing
images to look grainy, with very bright and very dark
pixels close to each other. It is caused by an object
which in a particular acquisition geometry behaves as
a strong reflector, or it is caused by the coherent sum
of the radar signals reflected by all objects in a certain
resolution cell, sum which randomly give a strong intensity.
5.
What are the main SAR platforms?
Several past, present and
future Earth Observation Satellites are SAR platforms
as well as the Shuttle Imaging Radar missions. See the
table for a full list.
| Sensor |
Launch Date |
| Seasat |
1978 - 1978 |
| SIR-C |
1994 - 1994 |
| ERS-1 |
1991 - 1995 |
| ERS-2 |
1995 - operational |
| JERS |
1992 - 1998 |
| Radarsat-1 |
1995 – operational |
| Envisat |
2002 - operational |
6.
What are the characteristics differencing the various
SAR systems?
The most important differences
between radar satellites are the angle with which the
microwaves are emitted (ESA-ERS satellites have a look
angle of 23°, Canada’s Radarsat has a varying look angle,
from 10° to 60°, Japan’s JERS has a look angle of 35°,
Envisat has a look angle of 14° up to 45°), the wavelength
of emission (C band for ERS and Envisat, L band for JERS,
C band for Radarsat) and the time elapsed between a passage
of the satellite and the following over the same area
(35 days for ERS, 44 days for JERS, 24 days for Radarsat
but the adjustable antenna can cover the same area approximately
every 3 days).
7.
What are the differences between radar and optical sensors?
SAR sensors, like any radar
sensor, can work without solar illumination and under
many weather conditions. They can be used at night as
well as during the day because they provide their own
illumination. Optical images are influenced by atmospheric
conditions (for example the presence of clouds, smoke,
aerosol or fog can turn an image useless); visible light
has short enough wavelengths to respond to all the individual
boundaries between air and water droplets. At each boundary
the light is reflected to a new direction, and by the
time it escapes the cloud, information on the light's
original direction is hopelessly lost. Radar signals can
instead penetrate clouds because the microwaves are not
subject to multiple scattering as visible light is. The
radar signals in fact are only affected while entering
and exiting the cloud. Because they don't suffer multiple
bounces, the radar waves are relatively undistorted by
clouds (from http://www.asf.alaska.edu/3_1.html).
8.
What is the difference between slant range and ground
range?
The SAR sensor emits first
a single pulse of microwave energy, then waits for the
returns from the ground. Energy reflected from objects
closer to the satellite (in the satellite-Earth direction)
will come back to the sensor before energy reflected from
objects further away. A certain delay in the return of
the signal corresponds to a certain distance from the
satellite. The receiver onboard the platform samples the
returned signal at high frequency. The time delay between
a signal and the following will correspond to a certain
relative distance between the two points from which the
signals originate. This distance is along the direction
the spacecraft is looking. The highest the frequency with
which the receiver samples the returns, the highest the
resolution in distance between two close points. As the
following image (from http://www.asf.alaska.edu/7_4_2_12.html) shows,
this resolution is called slant range resolution and is
about 8 meters for ERS. The slant range resolution projected
on the ground gives the ground range resolution, which
is about 20 meters for ERS.
9.
Why are SAR images called complex?
SAR images are called complex
images because they consist of complex numbers. The sensor
on board SAR platforms registers in fact the reflected
signal not only in terms of intensity but also in terms
of phase. The values of intensity and phase registered
are then expressed through a complex number.
10.
What is SAR data used for?
SAR's ability to pass relatively
unaffected through clouds, illuminate the Earth's surface
with its own signals, and precisely measure distances
makes it especially useful for the following applications:
- Sea ice monitoring
- Cartography
- Surface deformation detection
- Glacier monitoring
- Crop production forecasting
- Forest cover mapping
- Ocean wave spectra
- Urban planning
- Coastal surveillance (erosion)
- Monitoring disasters such as forest fires,
floods, volcanic eruptions, and oil spills (from http://www.asf.alaska.edu/3_1.html)
SAR INTERFEROMETRY
11.
What are the interferometric techniques and interferograms?
Interferometric techniques
are radar data processing techniques. Images acquired
by SAR satellites are matrices of values, each of them
regarding a certain point on the ground. Each value is
a complex number reporting the information in terms of
signal phase and intensity. To create the interferogram
two images are taken of the same area, acquired with two
different angles (in this case we talk of single-pass
interferometry), or at two different times (we talk of
repeat-pass interferometry). The difference between the
two phase values of the same point is calculated. An interferogram
is therefore a matrix whose values are the differences
between phases of a microwave signal. The usefulness of
an interferogram derives from the fact that this phase
differences are proportional to the object displacement
occurred in the time interval between the two acquisitions.
The phase difference is actually influenced by other effects
too, but if those can be removed, a precise displacement
measurement can be obtained.
12.
What are the output products of interferometry?
Interferometry produces in
output interferograms, which are contour maps of change
in the terrain-SAR platform distance. These maps can be
updated at every new passage of the satellite on the area
of interest (i.e. every 35 days for ERS), have high spatial
density (about 100 pixel/km2) and high precision. They
measure many effects: crustal movements, atmospheric perturbations,
dielectric properties changes of the terrain, topography.
They can be used to study terrain deformations or to create
DEM.
13.
Under which conditions can interferometry be applied?
Interferometry can be applied
when certain conditions are satisfied. For repeat pass
interferometry the distance between the two flight lines
should be no more than a certain value (about 1 km for
ERS). SAR satellites are designed to repeat their orbits
cyclically and generally this condition is satisfied.
Other conditions are a terrain slope not too steep, signal
wavelength long enough and a pixel resolution not too
coarse. The degree of steepness, the wavelength and resolution
acceptable depend on the distance between the two acquisitions.
A further condition that must be satisfied is that the
ground must be observed from the same direction in both
images. Last point is that in general it is not possible
to apply interferometry to couples of images belonging
to different satellites with different sensor characteristics
(for example with different wavelength or polarization).
Also, images of the same satellite but belonging to different
acquisition modes (ascending or descending) cannot be
used. In case of ERS-1 and ERS-2, since the two satellites
have identical instruments, their data can be used together.(adapted
from Massonnet and Feigl, 1998)
14.
What are the limitations of interferometry?
Interferograms are subject
to some limitations. The most influencing ones are temporal
and geometric decorrelation. The position and orientation
of objects on the ground can change in the time span between
the two acquisitions, resulting in a problem know as temporal
decorrelation, which limits the use of repeat-pass interferometry.
Also geometric decorrelation can limit it. It consists
of variations in reflectivity of the targets as a function
of the incidence angle. Interferograms are also affected
by two kinds of ambiguity. First, phase differences are
given in fractions of cycles (all pixels have phase between
0 and 1), not as integer numbers of cycles. Second, interferograms
provide relative phase changes, not absolute changes;
that means we must already know a point with null deformation
and refer all measurements to it.
High deformation gradients cannot be measured. The limit
is one interferometric fringe per pixel. There are also
dimensional limitations: interferometry must be applied
over many pixels, because when a single pixel is taken
in consideration, atmospheric effects and other noises
cannot be isolated. Moreover, it is necessary to study
geophysical phenomena that spread over at least 10 pixels.(adapted
from Massonnet and Feigl, 1998)
15.
What does baseline mean?
Two kinds of baseline exist
in SAR lexicon: normal baselines and temporal baselines.
Normal baseline is the perpendicular distance, usually
expressed in meters, between the positions of the SAR
platforms along their orbit at the two different acquisition
times. Temporal baseline is the time span, in days, between
the two acquisitions.
16.
How can interferograms be used in geophysics?
In order to correctly interpreter
interferograms as geophysical phenomena it is necessary
to be able to discriminate some artifacts which cannot
be removed and to keep in mind the limits of the techniques.
For example tropospheric effects still remain a source
of errors in interferometric measurements. They can tough
be identified because they produce fringes in all interferograms
including a certain date and disappear in interferograms
of other dates. Also topographic artifacts can be detected
because they appear in the same position in all images.
It is also important to evaluate the uncertainties in
measurements, by comparing them with other surveys (GPS,
etc).(adapted from Massonnet and Feigl, 1998)
17.
What kind of geophysical deformations can be detected
in an interferogram?
A geophysical phenomenon will
be visible in an interferogram if some conditions are
satisfied. The phenomenon must be larger than a single
pixel, but at the same time should be all within the radar
swath. The deformation rate should not be too high otherwise
the fringe pattern will become incoherent. This stands
also for abrupt changes in topography, like a surface
rupture. A deformation is significative only if it is
larger than the uncertainty in the measurement. If it
is not, it may be due to smooth atmospheric ramps of change,
or other effects. Last, the resolution in phase differences
cannot be higher than one tenth of a cycle, reducing therefore
the detection to signals larger than several millimeters
in case of ERS data and to centimeters in case of JERS.(adapted
from Massonnet and Feigl, 1998).
MULTI INTERFEROGRAM
TECHNIQUES AND PS TECHNIQUE
18.
What is the PS Technique?
The PS Technique (Permanent
Scatterers Technique) is an advanced satellite technology
which permits to measure very precisely movements (in
the Earth-satellite direction) of the ground surface using
radar images. Even movements of 0.1 millimeters per year
can be measured. This is possible because the phase of
the radar signal emitted by the satellite sensor and reflected
from the ground is proportional to the distance between
the satellite sensor and the point on the ground from
which the signal is reflected. The PS Technique was developed
at the Polytechnic University of Milan (POLIMI) and was
patented in 2000. It is the result of more than 10 years
of research of the SAR group at POLIMI.
19.
What does the PS Technique consist in?
The PS Technique consists
in a series of numerical elaborations of radar satellite
data, during which the data is studied statistically and
noise effects like disturbance due to the atmosphere,
errors in the reference DEM, etc. are removed. At the
end displacement values of points on the ground called
PS are obtained. To apply the Technique it is necessary
to have at least 25 – 30 radar images over the area of
interest.
In summary, the Technique
consists of the following steps. First of all images are
focalized. Focalization results in an increase of image
resolution. Then the radar images are registered to an
image chosen as master. Intensity values are normalized
to make the various acquisitions comparable. Then the
ratio between the average intensity and its standard deviation
over all images is calculated; when this value exceeds
a certain threshold (usually 2.5 or 3) it is indicative
of a candidate PS. All candidate PSs are called PSC (Permanent
Scatterers Candidates). At this point the differential
interferograms are created. Each of them is calculated
as the difference between an image and the master. Given
N images it is possible to create N -1 interferograms.
The value an interferogram assumes in each PSC point is
proportional to the PSC displacement plus a series of
other noises. Effects interfering with the displacement
value are: an atmospheric phase screen superimposed on
the image (called APS), errors in the reference DEM and
noise. Each effect can be removed considering its behavior:
for example atmospheric effects are correlated with space,
i.e. within short distances from a PSC they change proportionally
to space, but are uncorrelated with time and geometry
of acquisition, i.e. within small variations of time or
acquisition geometry they do not change proportionally.
Once noise effects are removed, a very precise measure
of the movements of each PS can be extracted.
20.
What are PSs?
The PS (Permanent Scatterers)
are points on the Earth surface that always keep the same
behavior in radar images acquired at different times.
With varying climatic and atmospheric conditions and different
satellite acquisition angles, they present the same electromagnetic
properties. They are usually parts of buildings, metallic
structures, rock outcrops, blocks of concrete. During
SAR image processing, PS are considered as lying in the
middle of the ground resolution cell (which has a size
of about 20*4 meters for ERS satellites). PS can have
varying dimensions; they can correspond to a TV antenna
on top of a roof as well as a large stone slab. In any
case they must have a point-wise behavior, therefore they
will always have small dimensions, far lower than the
resolution cell. The PS Technique allows the measurement
of millimetric displacements of these points. PS can be
interpreted as a network of natural GPS stations, with
a very high density and a monthly frequency (or higher,
depending on the frequency with which the radar satellite
acquire images) of update of the position information.
21.
What is the density at which PS are generally retrieved?
In urban areas PS are retrieved
with high density, even more than 500 PS/km2. Density
decreases in countryside (about 150 PS/km2) and is lowest
in mountain areas, where on average 20 PS/km2 are found.
In some cases of landslides, where the terrain is covered
with rocks and gravel, about 150 PS/km2 have been retrieved.
To apply the Technique, at least 4-5 PS/km2 must be retrieved,
otherwise it becomes impossible to correct the data for
atmospheric effects.
22.
What is the difference between the PS Technique and traditional
differential SAR interferometric techniques?
The PS Technique allows overcoming
some defects of the traditional interferometric techniques.
First of all, it produces estimates of displacement more
precise than the traditional techniques. Moreover traditional
interferometry suffers from various problems: temporal
decorrelation, geometric decorrelation, atmospheric influence.
The PS Technique removes atmospheric effects from the
estimation of movements using long time series of SAR
images. Temporal and geometric decorrelation are reduced
by choosing objects with point-wise behavior, that is
with reflectivity stable in time and at varying acquisition
geometries. Conventional differential interferometry can
be applied only with very small (< 200 meters) normal
baselines and on any number of images (? 2). The PS Technique
requires at least 15-20 images but baselines even higher
than the critical one can be used.
23.
How are, in detail, atmospheric effects removed from displacement
measurements?
The basic idea is to identify
different contributions (motion, topography, APS) to the
differential interferometric phase exploiting their different
behavior along three main dimensions: normal baseline,
time and space. The characteristics of each signal are
summarized in the following table:
| |
Time |
Space |
AcquisitionGeometry (Bn) |
| Motion |
Correlated |
variable |
uncorrelated |
| Topography |
uncorrelated |
variable |
proportional |
| APS |
uncorrelated |
correlated |
uncorrelated |
24. How and
at what stage of the processing are non linear movements
taken into account?
Non-linear movements are taken
into account only during the final phase of ground motion
estimation. The first steps of PS Technique are aimed
to Atmospheric Phase Screen (APS) evaluation and removal.
A subset of points called Permanent Scatterers Candidates
(PSC) have to be selected for this purpose, chosen among
the ones affected by linear deformation. This does not
represent a limitation, as the target here is not to extract
all the information available, but only a sparse grid
of points where the APS can be evaluated. Exploiting the
low-pass spatial behavior of the atmospheric disturbance,
the signal is then reconstructed over the whole image
and removed, allowing one to perform the final deformation
analysis. At this stage it is possible to use linear as
well as higher order model for motion.
25.
What do we mean by interferogram stacking?
Interferogram stacking is
an interferometric technique where the average of all
interferograms is calculated, in order to limit the impact
of atmospheric effects on the data. While in the PS Technique
the atmospheric contributions to the overall phase are
calculated and removed locally, in the interferogram stacking
technique they are removed based on the fact that considering
a large number of interferograms, the average atmospheric
effects tend to become null.
26.
Why are sometimes multiple PS retrieved, which are disposed
along a straight line?
Sometimes many PS are disposed
close to each other and along a straight line. This happens
because the radar signal reflected off the surface in
that area is very strong. Due to the signal filtering
and sampling, it results not as a unique large feature,
but the system represents it as a group of PS, where the
PS holding the highest coherence is the real PS and the
points at its side will have decreasing coherence and
will represent the same repeated information.
27.
What are the processing steps which are not performed
automatically but require an expert operator?
Most of the steps of the standard
PS Technique are performed automatically, but each of
them is always controlled by an expert. A few passages
are instead semi-automatic, as in the case of the master
image selection or reference point selection, in the sense
that the software proposes a choice but it is the operator
who has to accept it or make a different choice. In case
of the master selection, the image proposed by the system
has to be manually checked for meteorological conditions
before being accepted. The selection of PSC and PS is
performed automatically, so is the coherence threshold,
which is set so as to keep the probability of false detection
under 1:100,000. Check of the image quality is instead
completely carried out by the operator. In case of Advanced
PS Analysis the interaction with the operator is much
higher and longer computation times are required.
28.
Can the Technique be applied to steep relief areas?
Steep slope areas can give
problems in the application of the Technique as well as
in any interferometric analysis, because they might be
affected by layover or foreshortening or lie in a shadow
area. Some problems can be solved, as in the case of radar
shadows, where the choice of the right path and row of
the radar image can turn areas visible. Slopes steeper
than about 60° might lie in a direction orthogonal or
parallel to the radar signal direction, resulting in the
impossibility to determine the right location of points
on the slopes.
29. Which radar data is used
as basis for processing with the PS Technique?
Data from SAR sensors on board
radar satellites is used to measure movements and create
DEM. At present ERS-2 and Envisat satellites of the European
Space Agency and Radarsat-1 of the Canadian Space Agency
are in orbit. TRE uses this data and also images of the
previous satellites, which have already concluded their
missions (ERS-1 of the European Space Agency, operating
from 1991 to 1994 and JERS of Japan, active from 1992
to 1994). For further information on ERS and Envisat see
www.esa.int. For information on Radarsat satellites see
www.ccrs.nrcan.gc.ca. On the Japanese JERS see www.eorc.nasda.go.jp.
New SAR satellites are planned for the near future: Radarsat-2
and the Japanese ALOS. TRE is already working to adjust
the current processing algorithms to the characteristics
of the new satellites.
30. Can data
from different satellites be used together to create differential
interferograms and apply the PS Technique?
In general, every radar satellite
works on different wavelengths, frequencies, polarizations,
orbital paths, etc; it is therefore impossible to merge
data from different radar systems and use it to create
differential interferograms. In case of ERS though, ERS-1
and ERS-2 data can be merged, because the two sensors
are identical. The PS Technique can nevertheless be successfully
applied to datasets acquired by various satellites, provided
that the dataset consists of acquisitions of the same
sensor or of an identical sensor.
31.
What are the criteria for the selection of processable
images?
New radar images are acquired
continuously from many different satellites and thousands
of them are available in the archives, but not all of
them can be used for the PS Technique analysis. Images
are selected prior to purchase on the basis of their radiometric
quality as declared by the image provider. Customers may
suggest to discard certain images (such as in case of
processing of mountain areas where the customer knows
the surface was for example covered by snow on the acquisition
date). After focusing, SAR images are checked again and
discarded if they do not satisfy certain conditions. Once
the images are rejected, they cannot be recovered at a
later processing step.
Usually no more than 15 %
of the available images are discarded. The discard is
due to three main problems: high Doppler centroid value,
missing lines and high geometric baselines.
When the difference between
the Doppler centroid of an image and the Doppler centroid
of the master is too high, we talk of high Doppler centroid
values and we discard the image. Sometimes in the image
some lines are missing; in this case the whole image is
discarded. The geometric baseline is the distance in meters
between the two images. When there is a too large gap
between an image and its master, the images is discarded.
PRECISION
ASSESSMENT AND PRACTICAL PROBLEMS
32.
How much accurate are the PS movements measured?
The precision of surveys accomplished
by the PS Technique depends on the number of images available
over the area of interest, on the stability of each PS
and on the distance of the PS from the reference point.
This point is a point of known elevation and supposed
to be motionless. Precision of displacement measurements
is always between 0.1 and 0.5 millimetres per year in
case of average velocities. At a distance of less than
4 ~ 5 km from the GCP, accuracy of velocity can be 0.1
millimetres per year. In case of single measurements,
displacement accuracy is around 1 ~ 3 mm. The average
displacement velocity of each PS is provided to the customers
together with a quality index (called coherence of the
PS). The higher the coherence value, the higher the reliability
of the displacement measure.
33.
How is the accuracy of measurements related to the images
used?
Accuracy of the measurements
depends on the number of images used in the processing,
on their distribution in time and on the distribution
of the normal baselines. With increasing numbers of images,
accuracy increases. Given the same number of images though,
an evenly distribution of the acquisition dates and of
the normal baselines will result in higher accuracies
than an unevenly distribution, with time intervals without
data or a not Gaussian distribution of normal baselines.
34.
How much precise is the location of the PS points retrieved?
For each PS point retrieved,
a pair of coordinates is given (usually geographic latitude
and longitude or UTM northing and easting). Accuracy of
location of the PS depends on the resolution of the SAR
system, and it is around ±10 meters in East and ± 2 meters
in North. In order to give a PS its real coordinates,
that is in order to move from SAR coordinates (azimuth
and range) to ground coordinates, it is necessary to know
the PS elevation on the ground. The more precise the height
of the point, the highest the location accuracy. When
the elevation is not correct, the PS will be projected
on the ground in a wrong location.
35.
What kind of ground movements can be measured by the PS
Technique?
The PS Technique, like any
other interferometric technique, allows measuring movements
only along the satellite - Earth direction (called LOS
= line of sight), not in all directions. The following
figure shows what the LOS direction is. This means it
is not possible to detect pure horizontal or vertical
displacements, but of course, any movement along the LOS
can be projected and therefore can give vertical or horizontal
components.
Furthermore, it is possible to detect only slow movements
(roughly no more than 1 cm at every subsequent passage
of the satellite on the same area). This limitation is
due to the fact that what is really detected by the satellite
is the variation in the satellite – ground distance in
fractions of a wavelength. The PS Technique is therefore
optimal to detect subsidence or uplift phenomena, stability
of buildings and slow creeping landslides.
36.
What are the limitations of the PS Technique?
- The PS Technique is not able to detect
rapid movements (more than 1 centimeter every 35 days
in case of ERS), unless ground data are available reporting
the magnitude of displacement. This limitation is due
to the fact that SAR interferometry can measure variations
of the sensor-target distance equal to a fraction of
wavelength (equivalent to 5.66 centimeters for ERS satellites)
but when deformations are larger than an entire wavelength,
it is not possible to count the entire values but only
their decimal parts.
- Given the frequency of acquisitions
from satellite (every 35 days for ERS), it is not possible
to follow in real time the evolution of a phenomenon,
unless ad hoc acquisitions are planned (as in case of
Radarsat).
- Out of urban areas, on surfaces covered
by vegetation, the number of possible measurements is
sensibly reduced for lack of artificial structures or
rock outcrops. This limitation could be avoided installing
artificial benchmarks.
- It is necessary to have a dataset of
at least 15 -20 images of the same area to be able to
perform an analysis.
37.
Why is the precision of PS displacement velocities in
the order of millimetres and the precision of their heights
in the order of decimetres?
The precision with which displacement
velocity and heights are retrieved depends from the very
physical model explaining the phase values registered
by the satellite sensor. The phase of the microwave radar
signal registered by the sensor is the sum of many components:
a phase due to the elevation of the Earth surface, a phase
due to displacement movements and phases due to other
contributions. The phase due to the elevation of the surface
is linked to the difference in look angles between the
master and slave image and to the normal baseline between
the two images. The smallest detectable variations in
look angles and normal baselines result in measurements
of elevation in the order of decimetres. The phase due
to displacement movements is proportional to the distance
between the satellite and the ground and the smallest
detectable variations in this distance result in measurements
in the order of tenth of millimetres.
38.
Can the density of PS over a new site be predicted?
Before a new area is processed,
a feasibility study can be carried out, in order to understand
how many PS can be retrieved. The land cover, inclination
and direction the area looks in relation to the satellite
orbit can give an idea of the results. For example, slope
areas facing east and west can be easily detected by the
satellite signals because they are orthogonal to the satellite
path (satellites move in a north-south direction). Areas
covered by dense and continuous vegetation will not give
any PS. Landslides consisting of rocks or coarse gravel
result instead in many scatterer points. A feasibility
study can provide the location of all PS retrieved over
an area of interest. In order to carry out the study,
it is necessary to purchase all radar images and to process
them up to the step where the PS locations are found.
This procedure has a lower cost than the complete processing.
39.
Is there a way to increase the number of PS retrieved
over an area of interest?
In same cases, processing
only a subset of all available images, i.e. only the images
belonging to a certain time period, can result in an increase
of the PS retrieved over a certain area of interest. In
fact some radar targets have a point-wise behavior over
a certain time lap and therefore can be PS, but lose it
over longer time periods. As a drawback, working over
a shorter time period decreases the precision with which
PS displacements are measured and of course limits the
measurements to the time period analysed. Using overlapping
time spans might improve the number of PS found in each
time span but decreases the overall accuracy of displacement
measurements.
40.
Why should I resort to the PS Technique?
Through the PS algorithm it
is possible to obtain very precise ground displacement
measurements starting from 1992 (in case of ERS-1, later
for other satellites) up to date. There is no need to
carry a GPS to the location under study or to organize
optical leveling surveys. The PS analysis is therefore
less expensive and can provide results back in time.
41.
What are the applications of the PS Technique?
The PS Technique is suitable
for various applications.
- Detection of areas subject to subsidence
or uplift: Many projects have shown the presence of
ground deformation phenomena caused by pumping of
water, gas or hydrocarbons from the underground.
- Detection of land sliding areas and
instable slopes (over areas not covered with vegetation
and with sufficiently slow motion): In many cases
the PS Technique can provide quantitative information
often lacking in the analysis of active slopes. Satellite
data can also better detected the extension of instable
areas.
- Monitoring of volcanic zones and seismic
faults: Radar satellite monitoring can provide information
precious for civil defense.
- Planning of new roads and infrastructures:
Data obtained from the analysis of radar images can
be used to help taking strategic decisions on the
planning of new road tracks and infrastructures.
- Ascertain responsibilities in case
of damages caused by major works (subway lines, tunnels,
etc.): The possibility to have an historic archive
of radar data allows to verify the cause-effect links
between construction works and damages sustained by
buildings. The PS Technique is indeed a useful tool
for a retrospective analysis to solve judicial controversies.
It is of a great interest also for insurance companies.
- Stability check of private and public
buildings and architectural heritages: The PS Technique
can be used for monitoring and prevention, in order
to provide data used for precise and detailed controls.
In this case the PS algorithm can offer long time
series of measurements and reduced costs.
42. Does the
kind of satellite data used influence the results obtained
through the PS Technique?
Every satellite sensor emits
and receives radiation with a certain wavelength and polarization.
Also the angle under which the satellite looks at the
ground is different. The shortest the wavelength, the
highest the accuracy in the measurement of velocities.
This means that an analysis carried out using ERS data,
which works on C band (about 5 cm wavelength), will give
displacement velocities more accurate than a PS analysis
carried out using JERS data, which uses L band (about
23.5 cm wavelength). Polarization does not significantly
influence the results obtained through the PS Technique,
while the different look angles give different ground
resolutions; an higher ground resolution improves the
possibility to resolve PS. The look angle is linked to
the geometric distortions too, which can bring to PS being
hidden from the satellite view.
COMPARISON WITH STANDARD
TOOLS
43.
What is the difference between the PS Technique and GPS?
Displacements measured with
radar data and the PS Technique are almost vertical movements,
along the LOS (line between the satellite and the Earth).
GPS has little sensitivity to vertical movements, while
it detects with precision horizontal displacements (north-south),
which are instead hardly detectable with radar. Moreover
the PS Technique allows surveying PS movements with precision
10 times higher than GPS measurements and with a density
much higher. GPS has higher costs because it requires
an operator on the area of interest and can not provide
measurements back in time, unless they are available through
previous surveys.
Drawbacks of the PS Technique
as compared with GPS measurements are the fact that SAR
data are acquired not every day, but up to 44 days apart
(in case of JERS), while GPS networks provide daily data.
On the other hand, due to costs of station deployment,
installation and maintenance, fixed GPS networks working
for years are rarely found. Moreover, it is not an easy
task to identify the best station sites, to ask for proper
permissions, to prevent faulty operations, to protect
the area from spurious vibrations and electromagnetic
interference.
These features make the two systems somewhat complementary.
44.
What is the difference between the PS Technique and optical
levelling?
Optical levelling data can
be extremely accurate in vertical direction (up to a fraction
of millimetre), but errors are integrated while more and
more measurements are carried out. Thus the final accuracy
strongly depends on the number of benchmarks, their reciprocal
distances, and very often on the logistic conditions of
the surveying as well. Moreover 3-4 people are necessary
for accurate optical levelling surveys.
Optical levelling data are
much more precise in vertical direction with respect to
GPS surveys, but small horizontal displacements of the
benchmarks are usually not measured. In any case, it is
easy to compare PS data and optical surveying in a GIS
environment, possibly supposing that target motion is
merely vertical. On the other hand, whenever deformation
data are not full 3D, it will not be possible a quantitative
and rigorous analysis.
45.
How can PS Technique results be calibrated?
In general, both GPS and levelling
data can be successfully used to (1) calibrate the data
obtained applying the PS Technique, and (2) perform an
unbiased accuracy assessment of the final results. |
