Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Royal Military Academy -1- Centre Spatial de Liège
MUSAR:
A Multibeam Opportunistic
SAR System
25 May 11
E. Cristofani1, V. Kubica1, D. Derauw2
A. Orban2, C. Barbier2, X. Neyt1
1 Royal Military Academy, Signal and Image Centre, Brussels, Belgium
2 Centre Spatial de Liège, Université de Liège, Angleur, Belgium
Royal Military Academy -2- Centre Spatial de Liège
Contents
1. Introduction
2. Reception system
3. Exact focusing processing
4. RMA future work
5. Fast focusing processing
6. CSL future work
7. Application examples
Royal Military Academy -3- Centre Spatial de Liège
1. Introduction
• Objectives:
• To assess feasibility of bistatic SAR
imaging using C-band SAR satellites
• To assess bistatic vs. monostatic results
• To assess resolution and SNR
• Passive radar issues
• Lack of synchronization
Royal Military Academy -4- Centre Spatial de Liège
Typical Monostatic Scenario
Imaged area
Monostatic
configuration
Echoes
GOAL:
SAR
image
Tx - ground
Royal Military Academy -5- Centre Spatial de Liège
Bistatic Scenario
Imaged area
Bistatic
configuration
Direct path
Echoes
GOAL:
SAR
image
Tx - ground
Royal Military Academy -6- Centre Spatial de Liège
2. Reception System
AD
C
sref
Synthesized
image
Hardware Image synthesis
Matched
Filtering
+
Fast
processing
secho
Signal
Separation
Royal Military Academy -7- Centre Spatial de Liège
Hardware
AD
C
sref
Synthesized
image
Hardware Image synthesis
Matched
Filtering
+
Fast
processing
secho
Signal
Separation
Royal Military Academy -8- Centre Spatial de Liège
Hardware
Royal Military Academy -9- Centre Spatial de Liège
Location of the Receiver
Royal Military Academy
Optical view from the roof
Royal Military Academy -10- Centre Spatial de Liège
Acquired Signals
ENVISAT signals
A train of pulses
One LFM chirp (BB)
A satellite pass
Royal Military Academy -11- Centre Spatial de Liège
Signal Separation
AD
C
sref
Synthesized
image
Hardware Image synthesis
Matched
Filtering
+
Fast
processing
secho
Signal
Separation
Royal Military Academy -12- Centre Spatial de Liège
Signal Separation
• Typically, two antennas
• One antenna pointing to Tx, another one pointing to target
• Reference signal extraction (synchronization)
• By resynthesis
• Well-known transmitted signals (chirps)
• Extract parameters
• By spatial null-steering/beamforming (steering vectors)
• Steering vectors are unknown. Calibration is needed.
Royal Military Academy -13- Centre Spatial de Liège 25 Oct 10
Parameters Estimation
• α, φ and f0 extraction
Unwrapped
phase
Wrapped phase
Quadratic phase = LFM
Measured data
Synthesized chirp
Phase noise
Theoretical phase
Royal Military Academy -14- Centre Spatial de Liège
3. Exact Focusing Processing
AD
C
sref
Synthesized
image
Hardware Image synthesis
Matched
Filtering
+
Fast
processing
secho
Signal
Separation
Royal Military Academy -15- Centre Spatial de Liège
Exact Focusing Processing
SAR image
1. Acquired raw data 2. Range-compressed data
3. Azimuth-compressed
data
Matched Filtering
in range
Matched
Filtering
in azimuth
Early results
Royal Military Academy -16- Centre Spatial de Liège
Early Results
SAR image
• Solution: alternative scenario • Should contain water (lake, pond?)
• Using passive transponders?
Royal Military Academy -17- Centre Spatial de Liège
4. RMA Future Work
• Obtain longer acquisitions
• Assessing image resolution and SNR
• Compare bistatic vs. monostatic images
• Using point target (passive transponder)
Royal Military Academy -18- Centre Spatial de Liège
5. Fast Focusing Processing
Bistatic geometry
Raw signal modeling
Royal Military Academy -19- Centre Spatial de Liège
Fast Focusing Processing
• Considering geometrical peculiarities of
considered opportunistic SAR, some strong
approximations are possible.
• The general scheme may be simplified.
Royal Military Academy -20- Centre Spatial de Liège
General Considerations
We consider a fix receiver and
a moving emitter.
X is the azimuth coordinate of
the emitter.
rT is the range distance
between the emitter and
the considered point
target.
r0 is the minimum slant range
distance. x0 is the
corresponding azimuth
coordinate.
rR is the constant distance
between the point target
and the emitter.
Royal Military Academy -21- Centre Spatial de Liège
Geometrical Consideration
Data organization if considering focusing at constant range
gates:
Monostatic Bistatic
Translational symmetry
No Translational symmetry
“Data curvature”
Royal Military Academy -22- Centre Spatial de Liège
Fast Focusing Processing
• Two focusing strategies are possible :
• Interpolate raw data on a regular grid. Then focus
along constant minimum slant range
• Focus at constant range gates. Then
georeference and geoproject focused data.
• The second strategy is the easiest and the
most efficient.
Royal Military Academy -23- Centre Spatial de Liège
Fast Focusing Processing
• Fast Focusing Processing of the simulated data
provided by RMA:
➡ Simulated data:
• The receiver antenna is pointing toward the target.
• The receiver - point target distance is about 50m.
• The receiver is at an altitude 8 m higher than the point target.
• Both the direct signal (emitter - receiver) and the backscattered
signal (emitter - target - receiver) are given.
• Range focusing is implemented in the Fourier domain.
• Range focusing is performed using the direct signal.
➡ Either a single direct pulse is used for all azimuth positions or each
direct pulse is used at each azimuth position.
Royal Military Academy -24- Centre Spatial de Liège
Fast Focusing Processing
Range focusing of the simulated signal:
With constant replica
With azimuth-
dependant replicas
Range plot
Range plot
Azimuth
Ran
ge
Azimuth
Ran
ge
Royal Military Academy -25- Centre Spatial de Liège
Fast Focusing Processing
➡ The range-focused signal
may be seen as a
decomposition of plane
originating from a point at
azimuth position x0.
➡ Back-propagation of
these plane waves will
lead to the focused point.
Royal Military Academy -26- Centre Spatial de Liège
Fast Focusing Processing
• This back-propagation corresponds
mathematically to a Fourier transform.
• A simple scaled Fourier transform of the
range-focused data will lead to the image,
provided that range focusing was performed
using the direct signal and provided that the
observed points are close enough to the
receiver.
Royal Military Academy -27- Centre Spatial de Liège
Fast Focusing Processing
Focused simulated data:
Remark: The focused data is in the natural
acquisition geometry, i.e. on a curved grid.
Azimuth plot of focused
direct signal
Azimuth plot of focused
point target Range
Azi
muth
Royal Military Academy -28- Centre Spatial de Liège
6. CSL Future Work
• Verify the validity of the approximations
with respect to target-receiver relative
positions.
• Analyse the ability to translate and scale
the direct signal to extend the focusing
area.
• Test and validate the Fast Focusing
Process on real data.
Royal Military Academy -29- Centre Spatial de Liège
7. Application Examples
• Static and recurrent
observation of :
• Forests,
• Crops
• ...
• ...
Royal Military Academy -30- Centre Spatial de Liège
Application Examples
• Static and recurrent
observation of:
• Forests,
• Crops,
• Man-made structures,
• ...
• Interferometric
processing of recursive
acquisitions:
• Displacement/movement
monitoring.
Royal Military Academy -31- Centre Spatial de Liège
Application Examples
• If two reception
antennas:
• Direct and recursive
interferometric
measurement
• Coherence
monitoring for crop
stage/volumetric
estimation
• ...
Royal Military Academy -32- Centre Spatial de Liège
Contents
1. Introduction
2. Reception system
3. Exact focusing processing
4. RMA future work
5. Fast focusing processing
6. CSL future work
7. Application examples
Thank you! Any questions?