subpreg Documentation

subpreg - "Rubber sheet" sub-pixel registration of two images
SUBPREG runs a sub-pixel registration sequence on an image.  It runs
COREGPR over an image and creates a control file for WARP2.  The
sequence is as follows: COREGPR, WARP2 or RANDOM, and GEOM . For
further details read the help for each program in the sequence.

COREGPR
LTOL and STOL parameters are used by COREGPR as the maximum
misregistration that can occur in LINE and SAMPLE axes,
respectively.  LINE and SAMPLE is used to specify the starting line
and sample in the image to position the first box.  Sample position
remains a constant throughout processing.  NLBOX and NLBOX are
provided to allow for specification of the number of lines and
samples, respectively, in the box.  LINC and SINC specify the box
increment in lines and samples, respectively.  This may be any be
used to specify increments that result in overlapping boxes or
adjacent boxes (using LINC=0 and SINC=0).  To get a single box, set
BOTH LINC and SINC to values greater than the number of lines and
samples in the image.

WARP2
WARP2 is run after COREGPR using the control file as input.  It will
generate a tfile that is then used in GEOM to register the first input
image to the second.  If this option is selected in the FIT parameter,
RANDOM is not run.

RANDOM
RANDOM is run after COREGPR using the control file as input.  It will
generate a tfile that is then used in GEOM to register the first input
image to the second.  The DEG parameter is used to specify the degree
of equations to use to compute the registration.  If this option is
selected in the FIT parameter, WARP2 is not run.

GEOM
GEOM takes the first input file and the tfile generated by either
WARP2 or RANDOM and produces an image that is registered to the second
input file.

PROGRAMMER: Kris Becker, USGS, Flagstaff, AZ

Input image 1
Input image 2

NONE

Band number of image 1
and image 2

1,1

Output image

NONE

Control point file

"cfile.dat"

Transformation file for
 GEOM

"tfile.dat"

Maximum pixel misregistration
in line axis

5

Maximum pixel misregistration
in sample axis

5

Minimum correlation
      coefficient

0.9

Size of subarea in the line
axis

50

Size of subarea in the sample
axis

50

Box increment in line axis

0

Box increment in sample axis

0

RANDOM fit or
  WARP2 fit

"R"

Specifies the degree of
polynomial fit in random

1

Lowest dn value to use in
correlations

--

Highest dn value to use in
correlations

--

Minimum number of points
to use in best fit
calculation

5

Minimum fraction of box
to use in best fit
calculation

0.0

Starting line

1

Starting sample

1

Work disk

"./"

ADDITIONAL NOTES:

Parm	Description
FROM	The first input image will be registered to the second input image.
TO	File name of the output file. The output file is a registered image as a result of running through a subpixel registration processing sequence: COREGPR,WARP2 or RANDOM, and GEOM.
CFILE	CFILE is an ascii file that contains the control points that are input to the program WARP2. WARP2 requires a minimum of two points and a maximum of 200 points. Output of COREGPR. There is no default to avoid appending to the same output file for different input files.
TFILE	This is the transformation file that is needed by GEOM. Output of WARP2 and RANDOM.
LTOL	The maximum estimated pixel misregistration in the LINE axis. LTOL must be greater than zero. If LTOL is less than or equal to zero, COREGPR will assign LTOL=NLBOX.
LOW	The lowest value of valid DNs. If LOW = 11, then pixels with DN values lower than 11 will be excluded from processing.
HIGH	The highest value of valid DNs. If HIGH = 200, then values greater that 200 will be excluded from processing.
STOL	The maximum estimated pixel misregistration in the SAMPLE axis. STOL must be greater than zero. If STOL is less than or equal to zero, COREGPR will assign STOL=NSBOX.
MINR	Any control point that has a correlation coefficient (R) less than MINR, will not be written out to CFILE.
NLBOX	Size of the subarea box in lines that will be registered across the input images. Note that the box size is (NLBOX * NSBOX).
NSBOX	Size of the subarea box in samples that will be registered across the input images. Note that the box size is (NLBOX * NSBOX).
LINC	LINC is BOX INCREMENT for the LINE dimension in the image. The user should use wise judgement in setting this parameter. The effect of setting LINC=1 will offset each box by one line in the image, which could result in COREGPR executing a very long time. Letting LINC = 0 will result in the box being moved down the image by NLINES each time. This is probably the most common use. To compute only one point, set LINC to at least the total number of lines in the image. This will result in a single computation of boxes down the image. Note that the value for SINC will determine how many boxes across the image are computed.
SINC	SINC is BOX INCREMENT for the SAMPLE dimension in the image. The user should use wise judgement in setting this parameter. The effect of setting SINC=1 will offset each box by one sample in the image, which could result in COREGPR executing a very long time. Letting SINC = 0 will result in the box being moved across the image by NSAMPS each time. This is probably the most common use. To compute only one point, set SINC to at least the total number of samples in the image. This will result in a single computation of boxes across the image. Note that the value for LINC will determine how many boxes down the image are computed.
FIT	User has the option to run the control points through RANDOM or WARP2.
DEG	This parameter is only valid for RANDOM. It selects the degree of polynomial fit to apply to compute the transformation. See the RANDOM pdf for more details.
MINPTS	This parameter allows the user to select the minimum number of points used to calculate the best fit in COREPR.
FRAC	This parameter may be used instead of MINPTS to specify the minimum number of points to use in the calculation of the best fit. If this parameter is anything other than its default (0.0), then it will be used in COREGPR instead of MINPTS.
SL	Starting line. The default is 1 which will start the processing with the first line or top of the image.
SS	Starting sample. The default is 1 which will start the processing with the first sample or left side of the image.
WORK	Working disk for intermediate files that will be created by this procedure. These files are deleted at completion.