ScanSAR in InSAR applications

Document Sample
ScanSAR in InSAR applications Powered By Docstoc

        Jeremy Wurmlinger
        University of Kansas
            Spring 2009
 ScanSAR overview
 ScanSAR signal properties
 ScanSAR interferometry
 Spaceborne satellite ScanSAR
 Conclusion
 Questions/Comments?
                        ScanSAR overview
           Also known as “Burst mode” or “Wide Swath mode”
           Achieves a large swath coverage by periodically switching the antenna look
           Changing Antenna look angle results in have Multiple Swath
           During the look angle transfer, there is no transmit or receive(off period)
           Allows imaging of a wide swath area at the expense of azimuth resolution
           Typical swath widths range from 300km-500km
           Raw data consists of bursts of radar echoes that are shorter than the
            synthetic aperature length
           Various satellites are built with a ScanSAR mode: RADARSAT-1,
            RADARSAT-2, PALSAR among others.
      ScanSAR overview
             NB=Sequence of Burst Echo lines
             TB=Burst Duration
             TP=Burst Cycle Period
             T=Width of Antenna Footprint
             θ1=Look angle for first swath
             θ2=Look angle for second swath

             NL=T/TP (Burst Look: # of bursts per
                      synthetic aperture)

             NS=TP/TB (# of possible subswaths)
        ScanSAR Signal Properties
Impulse Response of a point scatter   Frequency Spectrum of a point scatterer
[2]     Strip-map mode                [2]          Strip-map mode

[2]       Burst mode                  [2]             Burst mode
        ScanSAR Signal Properties

      Top figure-shows the raw data of a single point scatterer in strip mode
      Bottom figure-shows the raw data of the same scatterer in 4-burst mode
      ScanSAR Signal Properties

         Time frequency diagram of a single scatterer located at t=t0
      WB=Burst Bandwidth
      WP=Distance between Spectral components of each individual burst
      WA=Chirp bandwidth
      HC(f;t0)=Signal Spectrum of the scatterer located at to
           ScanSAR Signal Properties

      Figure to the left depicts the impulse response in both strip and burst mode

      Figure to the right depicts the doppler centroid estimation of a scatterer in 2 Beam burst mode
                 ScanSAR Interferometry

           Multipass ScanSAR images may be used for interferogram
            generation to obtain elevation data
           Several issues present themselves in inteferometric ScanSAR
                 1. Critical Baseline and Bandwidth
                 1. Doppler Centroid Estimation
                 2. PRF Ambiguity Resolution
                 3. Azimuth Scanning Pattern Synchronization
                 4. Range Processing
                 5. Azimuth Proecssing
                 6. Interferogram Generation
      ScanSAR Interferometry

                   -Example of RADARSAT-1 datasheets
                    used for Interferometry
                   -Out of 12 RADARSAT datasets…two
                    pairs existed with baseline shorter
                    than critical baseline
                   -Out of the 2 datasets, only 1 pair
                    showed acceptable coherence

                   -For DTM purposes, this particular
                    ScanSAR dataset for RADARSAT
                    was able to acheive better resolution
                    than DTED-1
             Burst Mode Image Products
      Two Approaches to consider when approaching the phase-preserving data

      Single-Burst Image- Bursts are processed to individual complex images. The complex
         burst-mode product is then produced by these Burst complex images. Each image
         has a resolution to 1/WB and covers an azimuth extent of T-TB. One of the
         inconvenience asssociated with Single-Burst images are their azimuth-variant
         spectral bandpass characteristic. The signal varies linearly with azimuth from
         -WA/2 to WA/2 and wraps around several times in the +/- PRFn/2 band of the burst
         image This excludes the use of standard coregistration and resampling algorithms.

      Figure to the right shows the time frequency diagram
      of focused multiple burst RADARSAT ScanSAR data
            Burst Mode Image Products

      Multiple-Burst Image- involves the coherent superposition of the burst images forming a
         single full-aperture image. Suffers the multiple peak shape of figure in the right hand
         corner. It has been shown, however that a low-pass filter applied to the multiple-burst
         image gives a image comparable to the one obtained by incoherent superposition of
         burst images. The average spectral envelope of such a complex burst image is
         identical to that of a strip-map mode. Therefore, standard InSAR processing can be
         applied to multiple-burst images. The cost of this however is increased data volume.

                Critical Baseline and Bandwidth
        Critical Baselines for RADARSAT modes and ERS
                                                        -ScanSAR typically operates in a
                                                         reduced range chirp bandwidth
                                                        -Not only a resolution problem, but
                                                         also effects the size of the critical
                                                        -Core reason why Narrow beam mode
                                                         is preferred when using ScanSAR for
                                                         Interferometric purposes
                                               -A lower range chirp bandwidth also
                                                requires tighter margins for repeated
    (Critical Baseline increases with increase orbit trajectories
     in incidence angle)
[1] [5]
                                   Doppler Centroid
                                                      -Since the shape of the doppler spectrum
                                                       will be distorted in Burst mode….Doppler
                                                       Centroid estimation is more difficult than
                                                       Stripmap mode.

                                                     -Can be achieved by packing the bursts densely
                                                      in the azimuth direction, which would give an
                                                      uninterrupted sequence of range lines…then pass
                                                      through a Doppler Centroid Estimator(same
                                                      as strip mode)

          Yaw Steering- Antenna mid-beam axis is pointed in the direction of the
                        Zero Doppler frequency. This results in the majority of the
                        Doppler Centroid values being the 0th PRF band.
                        -Allows coregistration parameters to be derived for each individual
                       -Avoids Phase biases due to interferometric misregistration, PRF ambiguity
                        resolution,and Beam alignment and baseline estimate become much easier
                          PRF Ambiguity

      If a satellite in not yaw-steered, then PRF ambiguity resolution is necessary. Without
      the use of Yaw Steering….Doppler Centroid Values exceeding +-PRF/2 are expected

      Various methods can be applied to compensate for the PRF ambiguity. As discussed
      in Holzner and Bamler’s, a range look correlation technique is used. Since ScanSAR
      scenes cover areas about 10 times as large strip-map images, it is very likely that
      there are a few point targets in the scene. With the range look correlation, these
      point-like targets will be automatically detected. [1]

          The figure to the right shows the range
          Cross correlation of two looks of a point
          like target.
            Azimuth Scanning Pattern Synchronization
           Azimuth Scanning Pattern-azimuth pattern of the sequence of bursts(radar on) and
            pauses(radar off)
           Interferogram formation requires that the burst patterns of both datasets be aligned
            with each other
           The goal of Azimuth Scanning Pattern Synchronization is to perserve companions of
            raw data lines of the same ground area located in your image.
           poses a problem when trying to perform interferometric processing
           Another reason why narrow beam ScanSAR is preferred over wide beam ScanSAR.

        Figure to the right shows the variation of interferogram
        peak values vs mutual azimuth shift. The large peak
        indicates a high ASPS. The figure was formed from
        the method discussed on next slide.

        Azimuth Scanning Pattern Synchronization

        Pre-Processing Method [1-2]
        1.   Small azimuth portion of both image datasets is taken
        2.   Mutual azimuth shift is estimated from the two images
        3.   Co-registration of the two datasets(accuracy to 1 range line)
        4.   Only the range lines that related in for both datasets are kept for
             further processing; the rest are discarded.

             The more in sync the burst cycles of each dataset are with
             each other, the finer the final azimuth resolution of the
             interferometric image will be.

             “Mutually exclusive”-on cycles of one dataset line up with off
             cycles of the second dataset
                         Range Processing

        Two methods of Range Processing are but not limited to:
        -Single Burst Range Processing
        -Pack and Go Range Processing(proposed by Holzner and Balmer)

        Single Burst Range Processing-Each burst of range lines is processed
        separately by a standard processor whose azimuth compression has been
        disabled. This method is computationally efficient, however, often the bursts
        must be supplemented by a number of range lines containing zeroes to the
        next power of two. [1]
                         Range Processing
      Pack & Go Range Processing-Since during range processing (almost) no energy leakage
         in azimuth occurs, several bursts can be concatenated (with a few zero-lines as
         “safety margins” separating the different bursts) and processed simultaneously by the
         standard SAR processor.. This method is more efficient than single-burst processing,
         since arbitrarily long FFTs can be used and zero padding (to meet the power-of-two
         requirement) is negligible. Existing strip-map SAR processors can be used like a
         black box, no matter how large their internally used azimuth processing blocks are.
         Also Doppler centroid estimation that uses large estimation windows can be
         conveniently applied to this packed burst dataset. [1]

                         Azimuth Processing
      A focused burst image exhibits a range-dependent azimuth sample spacing known as fan-shape
      distortion. When performing Azimuth processing, an additional interpolation step is required.

      Azimuth Variant Filtering -After a burst has been range processed, standard azimuth compression is
      applied. The result is a coherent superposition of several shifted replicas of the desired burst image.A
      time-domain azimuth band-pass filter whose center frequency varies with azimuth position is required
      finally to separate the correct result from the aliased one.

      Extend Chirp Scaling- One method used to avoid the fan-shape distortion and the additional
      interpolation step could be done by using a high precision ECS algorithm, which would use the
      standard SPECAN method for the azimuth processing and the chirp scaling for accurate range
      processing. Adding a azimuth scaling step would also equalize the frequency rates of the azimuth

      Time-Domain Azimuth Correlation-Since the bursts are usually short, a straightforward time-domain
      azimuth correlation is an alternative to frequency domain methods. Although not as efficient, it is most
      easily implemented and allows for a free choice of the output sample spacing.
        Phase Preserving Burst Mode Processing

      Single Burst Images                      Multiple Burst Images
        -Burst Images acquired from each         -Multiple Burst images are stored
        pass                                     -Coherently add each individual burst
        -Individual Burst interferograms         image from each pass over the target
        formed                                   area.
        -Arrangement of Burst interferograms     -Interferogram formation
        to form subswath interferograms          -Low-Pass filter
        -Arrange subswath interferograms to      -NL interferogram is formed
        form NL interferograms

                   Final Interferogram formation essentially equivalent whether
                   from Single Burst method or Multiple Burst Method
      Single-Burst Interferogram Method

                               Companion Burst Images
                               acquired from each pass
                               Individual ‘Burst’ Interferograms
                               are formed

                               Subwath Inteferogram formed
                               by coherent addition of each
                               burst interferogram

                               Arrange subwath interferograms
                               to obtain the NL interferogram

      Multi-Burst Interferogram Method

                                Multiple Burst images are
                                stored for each Pass(N=3)

                                Coherent addition of each
                                individual burst image from
                                each pass to form a multiple
                                burst image

                                Interferogram is formed

                                Low-Pass filtering
                                NL interferogram is formed

        For the generation of subswath interferograms, the singleburst interferograms are
        coherently mosaicked. Straightforward multi-looking requires that the bursts are
        processed to the same grid within and across beams. Usually, the PRFs in the
        individual beams differ; hence, the raw data echo lines of the bursts within one swath
        are not acquired on a continuous integer pixel grid. The ScanSAR processor should
        arrange the data on a common reference grid to facilitate further processing.
        Differential interferograms generated between bursts within one subswath can be
        used to identify phase and sampling errors.

        Mutual registration of the interferometric partners requires interpolation of the signal
        at inter-pixel positions. For multiple and single-burst images sampled at the
        acquisition PRF, conventional interpolation methods can be applied since at the large
        PRF, the average burst signal spectrum resembles the one of strip-map data

        Depending on the sampling frequency, single-burst image re-sampling has to account
        for the azimuth variant band-pass behavior of the signal.

                         Technique proposed Holzner and Bamler
                         1. Burst Signal is de-ramped by a quadratic phase function in the
                          time domain such that a low pass signal is obtained
                         2. The low pas signal is then re-sampled using standard interpolation algorithm
                         3. The Quadratic phase is removed for further processing

         Alternative approach would be to apply azimuth-varying band pass for resampling.
                        Beam Alignment
        The swath interferograms can be either processed to individual DEMs or
        first combined to a full-swath interferogram and then processed to a single
        DEM. The swath overlap region provides valuable information on swath
        registration and mutual phase offsets of the interferograms.

        Swath co-registration can be verified by cross-correlation of homologous
        patches on an image contrast basis. Misalignment in range will not only
        disturb topographic information after swath interferogram combination but
        also generates a phase offset owing to the dominant (“flat earth”) fringe
        frequency of the swath interferograms. High Doppler centroid values,
        together with tiny misregistration of the interferometric partners will generate
        slowly varying phase errors that also affect the overlap area of the swath
                   Satellite Orbit Drift
Satellite Orbit Data, such as RADARSAT, can have low accuracy. ScanSAR
interferometry is optimal under certains conditions.

For RADARSAT-1 The westward drift lies within ±4 km, and the orbit boosts occur
approximately every 35 days. The minimum drift rate of the satellite occurs halftime
between boosts at the drift maximum. Acquisitions around this minimum drift rate are
favorable for ScanSAR interferometry since baselines are expected to be within the
critical baseline. [1]

Alternatively, Orbit Correction can be split into two tasks.
1. Correct the primary channel orbit for shifts in slant range and azimuth direction
2. Correcting the secondary channel optimize the baseline length and orbit angle.
            Satellite Orbit Drift

      ScanSAR interferograms
      ScanSAR interferograms
          RADARSAT-1 was launched at 14h22 on November 4, 1995 from Vandenberg AFB in California, into a sun-
          synchronous orbit(dawn-dusk) above the Earth with an altitude of 798 kilometers and inclination of 98.6 degrees.
          Developed under the management of the CSA in cooperation with Canadian provincial governments and the
          private sector, it provides images of the Earth for both scientific and commercial applications.

          NASA provided the Delta II rocket to launch RADARSAT-1 in exchange for access to its data. Estimates are that
          the project, excluding launch, cost $620 million (Canadian). The Canadian federal government contributed about
          $500 million, the four participating provinces (Quebec, Ontario, Saskatchewan, and BC) about $57 million, and the
          private sector about $63 million

          It C-Band satellite that operates at 5.3GHz and has the ability of a ScanSAR mode that provides potential swath
          widths up to 500km. While it used to achieve high quality images of the earth, it has the flexibility to support
          specific requirements.
          The RADARSAT-1 SAR utilizes beam forming techniques in order to provide several different
          beam modes. The modes may be capable of imaging closer to or farther from the satellite, with
          finer resolution and accordingly smaller areas or wide areas with worse resolution.

          RADARSAT-1 ScanSAR mode actually combines a few of these other modes in order to obtain a
          very wide swath - approximately 500 km, depending on which ScanSAR sub-mode is chosen.
          Three resolutions are provided: 75, 150, and 600 meters

                ScanSAR(Narrow mode)
                 -~250km nadir offset
                 -~400km nadir offset
                 -Swath width: 300km
                 -Range Resolution: 50m
                 -Azimuth Resolution: 50m
                 -Look(s): 2-4
                 -Incidence Angle: 20-40°,32-46° °

                ScanSAR(Wide mode)
                 -~250km nadir offset
                 -Swath width: 500km,440km
                 -Range Resolution: 100m
                 -Azimuth Resolution: 100m
                 -Look(s): 4-8
                 -Incidence Angle: 20-50°
       RADARSAT-2 was successfully launched on December 14 of 2007 for the CSA by Starsem, using a Soyuz FG
       launch vehicle. It was developed to be a follow-on to RADARSAT-1. It has the same orbit (798Km orbit) and is
       separated by half an orbit period from RADARSAT-1. Also operating in C-band, RADARSAT-2 was designed
       to support all existing RADARSAT-1modes and offers higher quality features ranging from improvement in
       resolution, to full flexibility in the selection of polarization options.

         Key Features and Benefits
         -3-100 meter resolution
         -HH,HV,VV,VH channel polarization
         -Left and Right Looking operation
         -Solid-State Recorders
         -On-Board GPS receivers
         -Yaw-Steering capable
         -Reduced programing lead time
         -Rapid tasking and priority programming levels
         -Improved processing abilities and speed

         -ScanSAR Wide-500km swath width
                      100m*100m resolution
                       20-49° incidence angle
                       Single or Dual polariztion

         -ScanSAR Narrow-300km swath width
                         50m*50m resolution
                         20-46° incidence angle
                         Single or Dual polarization
       RADARSAT-2 ScanSAR Wide Mode(VV,VH)
[14]         Brazil, March 3,2008
       RADARSAT-2 ScanSAR Narrow Mode(VV,VH)
[15]   Thunder Bay, Ontario, Canada March 3rd 2008
                        ALOS PALSAR
      The Advanced Land Observing Satellite (ALOS), launched on 24 January 2005, is a
      joint project between JAXA and the Japan Resources Observation System
      Organization (JAROS). ALOS has three remote-sensing instruments: the
      Panchromatic Remote-sensing Instrument for Stereo Mapping (PRISM) for digital
      elevation mapping, the Advanced Visible and Near Infrared Radiometer type 2
      (AVNIR-2) for precise land coverage observation, and the Phased Array type L-band
      Synthetic Aperture Radar (PALSAR) for day-and-night and all-weather land
                         ALOS PALSAR
[17]   With a wavelength of 23.6cm, PALSAR can obtain swath widths up to 350km while
       retaining a spatial resolution of 71-157m*100m(az) at 5 scan, short burst.
           Shuttle Radar Topography Mission(SRTM)

          In February of 2000, the space shuttle Endeavor carried a specially
           modified Radar system on a 11 day mission.
          Obtained elevation data on a near-global scale to generate the most
           complete high resolution digital topographic database of the earth.
          Only the use of ScanSAR interoferometry made it possible to map
           80% of the earth’s landmass.
       Shuttle Radar Topography Mission(SRTM)
      C-Band Antenna used Four Beam ScanSAR narrow mode to form 225km swath widths
      Illuminated two subswaths at a time using orthogonal polarizations
      By use of it’s polarimetric abilities, the ScanSAR duty cycle was 2:1 rather than 4:1
      The X-band antenna did not have steering cababilities and has to operate at a fixed nadir look
       angle of 52° and swath width of about 45km
      The X-band antenna was able to provide higher relative height accuracies by a factor of 2; but
       could not come close to the coverage that the C-Band ScanSAR was able to provide
    Shuttle Radar Topography Mission(SRTM)
   To obtain the DTM data the system was equipped with two antennas.
   One antenna was located inside the payload bay of the shuttle, while the other was
    located on one end of a 200ft mast that extended from the payload out into space.
   Terrain Data that was mapped was located between
    60 degrees North latitude and 54 degrees south latitude.
[18]   Shuttle Radar Topography Mission(SRTM)
       Shuttle Radar Topography Mission(SRTM)
                               SRTM main antenna

          -Consists of two antennas: X-band and C-band
          -C-Band Antenna used for the SRTM mission
          -Both Transmit and receive antenna
          -5.6cm wavelength
          -225km swath width
       Shuttle Radar Topography Mission(SRTM)
[18]                         SRTM outboard antenna

         -Consists of: X-Band and C-Band antennas, two GPS antennas,
         LED targets,and a corner cube reflector
         -Only receive antenna; main antenna used for transmit
         -mast was extended ~60m from the main antenna
The future of InSAR is highly dependent upon the ScanSAR interferometry and the
continued launch of satellites with ScanSAR capabilites. While sacrificing spatial
resolution for swath width is the cost of ScanSAR, future technology and processing
continues to grow and show finer resolutions with each new spaceborne satellite. The
ability to produce DEMs on a World wide coverage in a minimial amount of time is a
highly valuable resource. Which is one of the reasons why ScanSAR mode has
become a standard SAR mode on current and future spaceborne satellite missions.

As discussed, the ScanSAR interferometry introduces many complications that the
standard strip map mode does not have. The variations in the methods and
processes used to work around this complications are not defined and are up to the
Individual on how they would like to process that data. New algorithms and
processing techniques are being continuously to achieve the finer resolutions at the
higher swath widths.
[1]Bamler, Richard and Holzner, Jurgen, “Burst-Mode and ScanSAR Interferometry”, IEEE Transactions on
Geoscience and Remote Sensing, Vol 40, No. 9, September 2002
[2]Bamler, Richard and Holzner, Jurgen, “ScanSAR Interferometry for RADARSAT -2&3”, Can. J. Remote
Sensing, Vol 30, No. 3, pp. 437-447, 2004
[3]Bamler, Richard and Holzner, Jurgen, “RADARSAT ScanSAR Interferometry”, Geoscience and Remote
Sensing Symposium, 1999. IGARSS ’99 Proceedings, Vol 3, July 1999
[4]Guarnieri, Andrea Monti and Prati, Claudo, “ScanSAR Focusing and Interferometry”, IEEE Transactions on
Geoscience and Remote Sensing, Vol 34, No. 4, July 1996
[5]Jin, Michael, “Optimal Range and Doppler Centroid Estimation for a ScanSAR System”, IEEE Transactions on
Geoscience and Remote Sensing, Vol 34, No. 2, pp. 479-488, March 1996
[6]Tomiyasu,Kiyo,”Conceptual Performance of a Satellite Borne Wide Swath Synthetic Aperature Radar”, IEEE
Transactions on Geoscience and Remote Sensing, Vol GGE-19, No. 2, pp. 108-118, April 1981
[7]Shimada, M.A.R and Watanabe Manabe, “ALOS PALSAR: Technical Outline and mission concepts”, 4 th
International Symposium on Retrieval of Bio- and Geophysical Parameters from SAR Data for Land Applications”,
Innsbruck, Austria, Nov 16-19,2004
[8]Bamler, Richard,”The SRTM Mission: A World-Wide 30m Resolution DEM from SAR Interferometry in 11 Days”,
Photogrammetric Week ’99’, 1999

Shared By: