Gamma DIFF&GEO: Reference Manual

gc_map_fd

ANSI-C programs: gc_map_fd.c

NAME

gc_map_fd - Geocoding lookup table derivation based on DEM in MAP coordinates and SAR imaging geometry and table of Doppler parameters.

SYNOPSIS
gc_map_fd <MLI_par> <fdtab> <DEM_par> <DEM> <DEM_gc_par> <DEM_gc> <gc_map> [lat_ovr] [lon_ovr] [sim_sar] [u] [v] [inc] [psi] [pix] [ls_map] [frame] [ls_mode] [r_ovr]

<MLI_par>	(input) ISP MLI or SLC image parameter file (slant range geometry)
<fdtab>	(input) Table with Doppler polynomial parameters, in Gamma format
<DEM_par>	(input) DEM parameter file
<DEM>	(input) DEM data file (or constant height value)
<DEM_seg_par>	DEM segment parameters used for geocoding If this file exists, then the bounds of the DEM segment used for geocoding are read from the parameter file, otherwise the bounds are estimated using the SLC parameters and state vectors, and written to the new parameter file
<DEM_seg>	(output) segment of DEM used for geocoding interpolated if lat_ovr > 1 or lon_ovr > 1
<lookup_table>	(output) geocoding lookup table (containing range Doppler coordinates of DEM segment for geocoding
[lat_ovr]	latitude DEM oversampling factor (default = 2.0)
[lon_ovr]	longitude DEM oversampling factor (default = 2.0)
[sim_sar]	(output) simulated SAR backscatter image (in DEM geometry)
[u]	(output) zenith angle of surface normal vector n (angle between z and n)
[v]	(output) orientation angle of n (between x and projection of n in xy plane)
[inc]	(output) local incidence angle (between surface normal and look vector)
[psi]	(output) projection angle (between surface normal and image plane normal)
[pix]	(output) pixel area normalization factor
[ls_map]	(output) layover and shadow map (in map projection)
[frame]	number of DEM pixels to add around area covered by SAR image (default = 8)
[ls_mode]	output lookup table values in regions of layover, shadow, or DEM gaps: 0:set to (0.,0.) 1:linear interpolation across these regions 2:actual value (default) 3:nn-thinned
[r_ovr]	range oversampling factor for nn-thinned layover/shadow mode (ls_mode=3) (default=2.0)

Remark: enter - as filename to avoid creation of the corresponding output file

EXAMPLE
gc_map_fd 20070401.mli.par 20070401.fdtab srtm.par srtm.dem 20070401.dem.par 20070401.dem 20070401.lut 2 2 20070401.sim_sar u v linc psi pix ls_map

Creates geocoding lookup table based on DEM in MAP coordinates and SAR imaging geometry (oversampling factors in latitude and longitude 2.0, all DEM derived products are calculated and written out to files). The transformation between radar and map geometry is set up on the basis of the Doppler polynomial coefficients reported in the file 20070401.fdtab, which has been generated with the ISP program par_KC_PALSAR_slr.

DESCRIPTION
gc_map_fd calculates complex valued geocoding lookup table based on DEM in MAP coordinates, SAR imaging geometry and Doppler parameters contained in the fdtab file. With respect to the program gc_map, this program has the capability to use the doppler information stored in a table rather than using the doppler polynomials in the image parameter file. This approach is useful when the doppler changes are not well modelled using a low order polynomial. This program is useful in the case of ALOS PALSAR Wide-Beam data. The fdtab file is generated by the ISP program par_KC_PALSAR_slr. The geocoding lookup table has the dimension of the DEM segment in MAP projection and contains corresponding SAR coordinates as floating point complex valued numbers with the real part corresponding to the real-valued column and the imaginary part corresponding to the real-valued row number.

The program create_diff_par can then be used to create the DIFF/GEO parameter file used to measure offsets between a simulated SAR image and the actual image in order to correct the geocoding map produced by gc_map_fd.

The provided DEM can be in 4-byte floating or 2-byte (short) integer format and is characterized in the corresponding DEM parameter file.An image segment covering the area of the SAR image is extracted from the DEM and oversampled in range and azimuth using the user provided oversampling factors. This allows to generate geocoded products at user defined spatial resolution different from the available DEM.

In addition to the geocoding lookup table additional information is calculated from the DEM and SAR geometry. The quantities calculated are: local incidence angle, projection angle (for pixel size normalization), pixel size normalization factor, zenith and orientation angle for local surface normal, layover/shadow map. Output file names may be replaced by "-" to suppress creation of the corresponding output file

The layover and shadow mode flag ls_mode determines what values are written into the output lookup table for regions of layover or shadow. Mode 0 writes the values 0.,0. to the lookup table and corresponding areas geocoded using geocode_back will be set to the null value. Mode 1 functions like mode 0 with the exception that the gaps in the lookup table (corresponding to layover, shadow, or gaps in the DEM) are filled using linear interpolation from the lookup table values at the edges of the gap for each line.

Mode 2 writes the true values of the lookup table without consideration of layover or shadowing. This may lead to non-monotonic range coordinate sequences. Mode 3 is similar to mode 2 with the additional feature that samples of the lookup table that are in layover and are not very close to a sample in the SAR image are set to 0.0 (thinned out). In layover regions many points in the lookup table have very similar range pixel coordinates. The degree of thinning is controlled by the range oversampling factor [r_ovr] entered on the command line. This factor determines the minimum range spacing for adjacent samples in the lookup table. Large values of range oversampling lead to minimal thinning because the spacing between range samples is small. Use of mode 3 introduces small gaps in the terrain corrected output products for layover regions in the SAR image. In modes 0, 2,and 3 gaps in the DEM are set to (0.0,0.0) in the lookup table.

Notice that the geocoding lookup table calculated is based on the available DEM and SAR geometry information. In particular the orbit data is usually not perfectly accurate, so that we recommend including a fine registration step as described in the DIFF Users Guide. The geocoding lookup table is used by the programs geocode and geocode_back for forward and backward geocoding. The geocoding lookup table may be inverted using the program geocode_back.

The simulated SAR intensity image can be used for the fine registration step. For this purpose it is first converted to SAR range/Doppler coordinates (using the program geocode). Then, in a cross-correlation analysis with a real SAR image, determine the fine registration offsets and use these to refine the geocoding lookup table. An alternate method for converting the simulated SAR image into SAR coordinates is to first invert the geocoding lookup table using gc_map_inversion and using geocode_back to resample the simulated image into SAR range/Doppler coordinates.

OPTIONS
The following DEM projections are supported:

Keyword	Projection
EQA	Equiangular
UTM	Universal Transverse Mercator
TM	Transverse Mercator (Germany, Finland, Italy, and others)
OMCH	Oblique Mercator (Switzerland)
LCC	Lambert Conformal Conic (France)
PS	Polar Stereographic
SCH	SAR Coordinates used at JPL
PC	Polyconic Projection
AEAC	Albers Equal Area Conic

For the generation of Geocoded Ellipsoid Corrected (GEC) products a constant height value can be specified instead of the input DEM file. GEC geocoding does not require a DEM. In spite of the non-existence of an input DEM and the constant height value used a DEM parameter file has to be created in order to specify the desired projection, ellipsoid and pixel spacing.