itg stereo matching, camera modelling and Kriging pipeline harness
-usage
displays short form usage string.
-version
displays program version number.
-slots
displays short form
pups(3)
library dependencies.
-slotinfo
displays long form
pups(3)
library dependencies.
-argf
argument_file
tells the filter to take its command line arguments from the specified
argument_file.
The data format of the argument file is:
# ... optional comment line ...
-argflag [1]
argument_parameter [1,1]
argument_parameter [1,2]
argument_parameter [1,n]
# ... optional comment line ...
-argflag [2]
argument_parameter [1,1]
argument_parameter [1,2]
argument_parameter [1,n]
-nice
cardinal
determines the
niceness
at which the filter is scheduled. In keeping with all UNIX user processes, nice
must be a cardinal between 0 and 20. The default niceness for
itg(1)
is 4.
-delete_tmpfiles
tells
itg(1)
to delete temporary files as soon as the stereo system has finished with them. Defaults
to TRUE (delete temporary files).
-image_names
left_image_name
right_image name
specifies the names of the (stereo) images to be matched.
-pg_leader
tells
itg(1)
to become the session (process group) leader for the stereo matching pipeline using
setsid(2).
This means that any signals sent to
itg(1)
are automatically passed to the pipelines which it sets up. This is useful when a stereo matching job
run amok - killing the harness will automatically kill all processes associated with the stereo matching
job
-test
tells
itg(1)
to print the shell scripts it generates on
stderr
rather than executing them. This option is intended for debugging
-delete_raw_stereo_images
tells
itg(1)
to delete the raw stereo images after the
prc(1)
pipeline has been run. This option only has an effect if
itg(1)
is pre-processing the input imagery.
-display_coverage
coverage_basename
tells
itg(1)
to invoke
coverage(1)
in order to generate a pair of coverage. In order to conserve machine resources, the
basal tier imagery of the autoseeding pyramid is used as the backdrop imagery for the
coverage plots. Typically, this means coverage will be scaled by a factor of 10 (for
SPOT imagery) relative to the DDM and DTM. The
coverage_basename
is used to generate
ref.coverage_basename
and
nref.coverage_basename
which contain the coverage relative to reference and non-reference images respectively.
-display_progress
tells
itg(1)
to invoke
Xhips(1)
to interactively display the progress of the DTM creation algorithms relative to the
image
image_name
(
ref
or
nref
).
Defaults to the non-reference image
(
nref
). In addition an
Xhips(1)
viewer is popped up at the end of the stereomatching process to display
the Kriged DTM in byte image form. This byte image is produced from the Kriger float HIPS
output image and is automatically subsampled to about 600 pixels square in size. After the
viewer has read its data it detaches itself from the pipeline that spawned it and sets up
its own process group. This means that even if
itg(1)
fails to complete any viewers which it created will still be on the desktop and can therefore
be used to investigate the problem.
-seeds
tells
itg(1)
to pop up an
xhips(1)
viewer to display the seed points generated. Defaults to TRUE
(seed display popped up). The
ref/nref
image option (to select the image onto which
progress O/P is plotted) is required for this option.
-matches
tells
itg(1)
to pop up an
xhips(1)
viewer to display the progress of the stereomatcher. Defaults to FALSE
(No stereomatcher progress display as it slows the pipeline down appreciably if
the X display is slow). The
ref/nref
image option (to select the image onto which
progress O/P is plotted) is required for this option.
-dtm
tells
itg(1)
to pop up an
xhips(1)
viewer to display the DTM generated (at reduced resolution). Defaults to
TRUE (DTM display popped up). If this progress display option is selected,
the
ref/nref
output image option is not required.
-seed_image_base_size
sets the size of the image which will be used as the basal image for the
autoseeder resolution pyramid. Defaults to 600 (which appears to work well
with SPOT-PA imagery).
-seed_image_base_scale
scaling factor
tells
itg(1)
the factor by which the input imagery is rescaled (by the
rescale(1)
image rescaler) prior to being passed to the
cascade(1)
autoseeder. Thus, the basal tier of the resolution pyramid processed by
cascade(1)
is scaled relative to the input imagery by a factor
scaling_factor.
The seeds produced by
cascade(1)
are subsequently adjusted to the correct scale for the input imagery by
pslam(1).
If no scaling factor is given the fitler autoscales so the basal tier of the
seed image pyramid is 600 pixels square.
-cascade_agf
cascade autoseeder argument file
is used to pass the argument file for the
cascade(1)
autoseeder to the autoseeding pipeline. Details of the options for
cascade(1)
are given in the
cascade (1)
man
page.
-pslam_agf
lam stereo matcher argument file
is used to pass the argument file for the
pslam(1)
stereo matcher to the matchtran pipeline. Details of the options for
pslam(1)
are given in the
pslam (1)
man
page.
-camera_model
camera model filter name
specifies the camera model which is to be used to transform stereo conjugate points
to ground truth. The camera model argument file, co-ordinate conversion and Kriging
options are only valid if a camera model has been selected.
-camera_model_agf
camera model argument file
is used to pass the argument file for the camera model. Details of the
options for the camera model are given in the relevant camera model
man
page.
-cconv
co-ordinate conversion filter
specifies the co-ordinate conversionfilter which is to be used to transform points produced by
the camera model into an appropriate local vertical system (such as UTM, Lambert Zone 3, UK National
Grid etc). This option is intended to be used when terrain images are being matched. For imagery
aquired in a vision cell co-ordinate conversion is not generally required.
-cconv_agf
co-ordinate conversion filter argument file
is used to pass the argument file for the co-ordinate conversion filter. Details of the
options for the co-ordinate conversion filter are given in the relevant
man
pages.
-krige_agf
I Kriger argument file
tells the
itg(1)
filter to set up a Kriging pipeline. This takes the ungridded DTM produced by the
matchtran pipeline and interpolates onto a regular grid using the
krige_spb(1)
fast Kriging algorithm. The options for the fast Kriging algorithm are given in the
krige_spb(1)
man
page.
-krige_dtm_name
gridded DTM output file name
tells the Kriger where to write gridded O/P data. If the
-krige_agf
option is selected, the DTM O/P file must be specified.
-ugrid_dtm_name
ungridded DTM file name
specifies the name of the file in which ungridded DTM is stored. This option is only available
if the
krige
option has
not
been selected.
-match_grid
N_x_points N_y_points
tells
itg(1)
to compute compatible set of
pslam(1)
grid parameters and
krige_spb(1)
bin parameters. If incompatible sets of grid/bin parameters are specified (with more Kriger bins than there
are points on match grid of the stereo matcher) undesirable artifacts may appear in the O/P DTM. If the
-match_grid
parameter is set the
-grid
and
-bin_size
parameters in the
pslam(1)
and
krige_spb(1)
argument files must not be set (or an error will result).
-match_image_scale
scale of match image
specifies the scale of the match image (relative to the image acquired by the camera model). This option
allows the user to generate DTM's from imagery which has been rescaled after capture and prior to stereo matching.
-ref match_image_offset
X offset
Y offset
specifies the (reference) image offset (added after stereo matching) to the stereo matched points. This option permits the
user to generate DTM's from imagery which has been extracted (as a region of interest) from the original
images after capture and prior to stereo matching.
-nref match_image_offset
X offset
Y offset
specifies the (non reference) image offset (added after stereo matching) to the stereo matched points. This option permits the
user to generate DTM's from imagery which has been extracted (as a region of interest) from the original
images after capture and prior to stereo matching.
-xyz
tells
itg(1)
to produce a kriging pipeline which write DTM grid data (via a LZW compression filter) to the
specified DTM file (X,Y,Z format).
-xyzv tells itg(1) to produce a Kriging pipeline which pipes DTM data (via a LZW compression filter) to the specified DTM file (X,Y,Z,V format). br
-no_proc
Tells
itg(1)
not to pre-process imagery using the partial rank correlation
prc(1)
filter.
-prc_patchrad
sets the patch radius parameter which is passed to the
prc(1)
filter
for both the
autoseeding
and
matchtran
phases of stereo matching.
-ddm_name
ddm file name
tells
itg(1)
to store stereo matcher O/P is ASCII format in
DDM file name.
SH SIGNALS
SIGUSR2 is used to support PSRP event handling