Gamma ISP: Reference Manual


sbi_filt.c

ANSI-C program: sbi_filt.c

NAME
sbi_filt  Azimuth filtering of SLC data to support split-beam interferometry to measure azimuth offsets

SYNOPSIS
sbi_filt <SLC-1> <SLC1_par> <SLC2R_par> <SLCf> <SLCf_par> <SLCb> <SLCb_par> <norm_sq> [iwflg]

<SLC-1> (input) SLC-1, reference geometry (fcomplex or scomplex format)
<SLC1_par> (input) SLC parameter file of SLC-1
<SLC2R_par> (input) SLC2 ISP image parameter file for the co-registered image of the interferometric pair, used to determine azimuth common-band for each output SLC (enter - for none)
<SLCf> (output) single-look complex image (forward-looking, fcomplex format)
<SLCf_par> (output) SLC parameter file (forward-looking)
<SLCb> (output) single-look complex image (backward-looking, fcomplex format)
<SLCb_par> (output) SLC parameter file (backward-looking)
<norm_sq> squint between beams as a fraction of the azimuth spectrum width (default: 0.5)
[iwflg] inverse weighting flag:
              0: no compensation for azimuth spectrum weighting
              1: compensate for the azimuth spectrum weighting (default)

EXAMPLE

sbi_filt 19991020.rslc 19991020.rslc.par 19991020.rslc.par 19991020f.rslc 19991020f.rslc.par 19991020b.rslc 19991020b.rslc.par .5  1

The input SLC image is filtered to produce 2 SLC images, 19991020f.rslc is squinted forward and 19991020b.rslc is squinted backward.

DESCRIPTION

sbi_filt is used to filter single look complex (SLC) SAR images, creating two SLC images with different squint angles, one looking foreward and the other backward.  An interferometric pair of SLC images processed in this way can then be used to create two interferograms, with the first interferogram from the forward-squinted beam and the second from the backward-squinted beam. The difference in phase of these two interferograms can be shown to be related to the along-track (azimuthal) displacement as described by Bechor and Zebker (2006).

The total azimuth spectral width of the SLC is determined by the radar along-track sample rate called the pulse repetition frequency or PRF.  The squint of the beam is determined by the radar orientation and latitude. The expression for the doppler centroid as a function of squint angle relative to look vector perpendicular to the velocity is given by:

Squinting the beam away from broadside towards the direction of travel gives a positive doppler, away from the direction of travel gives a negative doppler shift. The frequency separation in doppler between the looks determines the angular separation between the two synthesized beams.

The center of the SLC azimuth spectrum is called the doppler centroid and can vary both along-track and perpendicular to the track due to changes in the satellite latitude and orientation during an orbit. Filtering of the SLC data is performed using a pair of finite impulse response (linear phase) filters and applied using frequency domain convolution. A different pair of filters is calculated for  every 256 range samples across the track and every 2000 lines along-track. These filters track the variation in the doppler centroid determined from the doppler centroid polynomials in the SLC parameter file.

The optimum separation between the synthesized beams is approximately 0.6 times the antenna beamwidth. This is equivalent to a difference of the doppler centroid of the two beams of 1/2 PRF.The filters are specified by the center frequency and bandwidth. The norm_sq command line parameter specifies the frequency separation of the center frequencies of the two filters as a fraction of the PRF. The filter bandwidth is calculated such that it is the maximum possible within the  processing bandwidth and is symmetrical with respect to the chosen center frequency. The processing bandwidth of the SLC is the fraction of the PRF bandwidth used to produce the SLC image. Usually the processing bandwidth is in the range of 0.8 to 0.9 times the PRF in order to reduce azimuth ambiguities.

An azimuth windowing function is usually applied to the SAR radar data in order to reduce  azimuthal sidelobes. The program has the option to apply an inverse weighting function of the spectrum that increases the available bandwidth and resolution. Due to the reduced bandwidth of the squinted images, this option significantly improves the interferometric correlation of the interferograms produced from the split-beam SLC data.

The output SLC data are in floating point complex (fcomplex) format, even if the input SLC images are short complex (scomplex) format. The output filtered SLC data are exactly coregistered with the input SLC. The doppler polynomial information is updated to reflect the new doppler centroid of the filtered data. This information is used by sbi_offset to convert the interferometric phase of the split-beam interferogram into along-track displacement.

Bechor, N., H. Zebker,  "Measuring two-dimensional movements using a single INSAR pair", Geophsical Resarch Letters, Vol. 33, L16133, 2006, doi:10.1029/2006GL026883

SEE ALSO
typedef_ISP.h, SLC_par, OFF_par, sbi_offset


© Copyrights for Documentation, Users Guide and Reference Manual by Gamma Remote Sensing, 2011.
UW, CW, last change  25-Nov-2011.