ANSI-C program: base_init.c
NAME
base_init - initial estimation of interferometric baseline
SYNOPSIS
base_init <SLC1_par> <SLC2_par>
<OFF_par> <int> <base> [mflag] [nrfft] [nazfft]
[r_samp] [az_line]
<SLC1_par> | (input) SLC-1 Parameter File (of reference pass) |
<SLC2_par> | (input) SLC-2 Parameter File (of slave pass) |
<OFF_par> | (input) ISP Processing/Offset Parameter File |
<int> | (input) multi-look interferogram with range phase |
<base> | (output) baseline parameter file |
[mflag] | flag to select baseline estimation method (default=2) |
[nrfft] | size of range FFT (512,1024,...) |
[nazfft] | size of azimuth FFT (512,1024,...) |
[r_samp] | range pixel offset to center of the FFT window (default=center) |
[az_line] | line offset from start of the interf. for the FFT window (default=center) |
EXAMPLE
base_init 1352.slc.par 1610.slc.par 1352_1610.off
1352_1610.int 1352_1610.base 2
The baseline component parallel to the look vector is estimated based on the orbit parameters in 1352.slc.par and 1610.slc.par, the baseline component perpendicular to the look vector is estimated based on the interferogram fringe frequency (using fft techniques).
base_init 1352.slc.par 1610.slc.par - - 1352_1610.base 0
The baseline is estimated based on the orbit parameters in 1352.slc.par and 1610.slc.par.
base_init 1352.slc.par 1610.slc.par 1352_1610.off - 1352_1610.base 1
The baseline is estimated based on the orbit parameters on the SLC registration offsets, respectively based on the range and azimuth pixel spacing scale factors in the case of slc which were registered to a common geometry.
base_init 1352.slc.par - - 1352_1610.int 1352_1610.base 4
The baseline component parallel to the look vector is not estimated (set to 0.0), the baseline component perpendicular to the look vector is estimated based on the interferogram fringe frequency (using fft techniques).
DESCRIPTION
base_init estimates the interferometric baseline.
The estimated baseline and (along track) baseline rate, in TCN coordinates, is written to the first two lines of the baseline file. The baseline components parallel and perpendicular to the look vector are written to stdout.
In order to achieve on one hand a high applicability and on
the other hand high quality estimates several baseline estimation
algortihms are included. The included algorithms allow to
estimate the interferometric baseline and the (along track) rate
of change of the interferometric baseline from:
- the orbit state vectors
- the SLC registration offsets
- the interferogram fringe rate
In practice it turns out that it is usually best to combine two of these methods. The first two methods give the better estimates for the parallel baseline component. The third method gives very often the better estimate for the perpendicular baseline component. Therefore, it is recommended to combine either one of the first two methods with the third method.
In order to allow this the user can select between the individual methods and combined methods as listed in the table below.
Method flag | Bpara estimation | Bperp estimation | required input |
0 | from orbits | from orbits | SLC-1_par,SLC-2_par |
1 | from registration offsets | from registration offsets | SLC-1_par,SLC-2_par,OFF_par |
2 | from orbits | fringe rate | SLC-1_par,SLC-2_par,OFF_par,interferogram |
3 | from registration offsets | fringe rate | SLC-1_par,SLC-2_par,OFF_par,interferogram |
4 | fringe rate | fringe rate | SLC-1_par,OFF_par,interferogram |
Names for input files which are not required for a selected method can be replaced by a - sign. The program checks if the selected method is possible with the provided data. If that's not the case it automatically selects an alternative method which for which the required input data is provided.
As an option the user can select the range and azimuth fft sizes used for the determination of the fringe frequency. Notice, that the fringe frequency estimation is done iteratively to ensure maximum reliability of the perpendicular baseline component which is so important in SAR interferometry.
A large range fft size will reduce the accuracy of the first Bperp estimation because of its slant range dependence. In the iteration the interferogram is flattened with the prior baseline estimate which will remove this problem, in particular if a relatively accurate Bperp estimate is used. As a consequence the combined methods with such relatively accurate Bperp estimates (originating from an other method) lead to the best results.
In the fringe rate method the assumption that the surface on the average flat for the fft window is used. If that's not the case the Bperp estimation will be slightly wrong and there will be a residual phase trend in a flattened interferogram. If the user knows that the surface is flat for a certain part of the scene he may select this location for the fft baseline estimation. The baseline parameter file will be used for the removal of the average phase trend of the interferogram and for the computation of height and slope. The initial baseline estimation can be improved based on ground control points using the program base_ls.
OPTIONS
none
NOTES
The program base_init replaces the programs base_orbit
and base_est_fft of earlier ISP versions. The capability
of these older baseline estimation programs has been fully
integrated into base_init.
SEE ALSO
baseline_file, base_ls, typedef_ISP.h.
© Copyrights for Documentation, Users Guide and Reference Manual by Gamma Remote Sensing, 2000.
UW, CW, last change 24-Jan-2000.