iraf.net

Dear IRAF staff,I have some images with WCS in the headers. I also have a list of RA DEC coordinates of stars that could be used for photometry. I want to run some IRAF task, I guess ccxymatch, in order to transform the RA DEC to pixel coordinates. The images are all of the same field, but have been dithered with large offsets of about 1arcmin. I have no tie points and since i intend this to be in the middle of a pipeline I can't click to obtain tie points either, hence I want to use the triangles algorithm.
For some images the ccxymatch task finds some matches, say 4 or more. For others, not so very different looking images, ccxymatch finds nothing. For which images it fails and for which it succeeds seems to be very arbitrary?? Perhaps maybe it depends on:1. how long the input RA DEC list of star coordinates is
2. the order of the RA DEC star coordinates in the input list??
3. the max ratio of the triangle sides
My feeling is that the performance of the triangle algorithm is rather arbitrary and can't really be controlled. Is that so? What should I tweak to make it find the right triangles to match? To illustrate, for these star coordinatesStar2 9.200208 -33.534256
Star3 9.200566 -33.636135
Star4 9.214608 -33.587601
Star5 9.196777 -33.512306
Star9 9.202347 -33.555851
Star10 9.201588 -33.593369
Star11 9.126136 -33.499496
Star15 9.220608 -33.497650
Star17 9.285759 -33.505417
Star19 9.290563 -33.608166
Star20 9.217588 -33.648449
Star22 9.282237 -33.543961
Star23 9.157138 -33.548901
Star24 9.297270 -33.489647
Star25 9.288115 -33.563831
Star26 9.234565 -33.646046
Star28 9.187657 -33.581608
Star29 9.162666 -33.556934
Star30 9.285396 -33.482327
Star32 9.242652 -33.622772
Star33 9.180610 -33.540237
Star34 9.245084 -33.548298
Star35 9.101016 -33.564934
Star37 9.242251 -33.503075
Star38 9.305925 -33.592663
Star39 9.310121 -33.614285
Star40 9.171402 -33.631947
Star42 9.261161 -33.560287
Star43 9.205060 -33.552872
Star44 9.217359 -33.595188
Star45 9.145441 -33.478840
ccxymatch DOES FIND stars from this header:[b:87d4a23657]RA = ' 00:36:50.419' / 00:36:50.4 RA (J2000) pointing
DEC = '-33:34:22.584' / -33:34:22.5 DEC (J2000) pointing
EQUINOX = 2000. / Standard FK5 (years)
RADECSYS= 'FK5 ' / Coordinate reference frame
CRVAL1 = 9.21008 / 00:36:50.4, RA at ref pixel
CRVAL2 = -33.57294 / -33:34:22.5, DEC at ref pixel
CRPIX1 = 488. / Ref pixel in X
CRPIX2 = 488. / Ref pixel in Y
CDELT1 = -0.00008000 / SS arcsec per pixel in RA
CDELT2 = 0.00008000 / SS arcsec per pixel in DEC
CTYPE1 = 'RA---TAN' / pixel coordinate system
CTYPE2 = 'DEC--TAN' / pixel coordinate system
PC1_1 = 0.000000 / Translation matrix element
PC1_2 = 1.000000 / Translation matrix element
PC2_1 = -1.000000 / Translation matrix element
PC2_2 = 0.000000 / Translation matrix element
EPOCH = 2000.0
JD = 2453653.76390995
HJD = 2453653.76827793
LJD = 2453653.
WCSDIM = 2
CD1_1 = -8.0000000000000E-5
CD2_2 = 8.00000000000000E-5
LTM1_1 = 1.
LTM2_2 = 1.
WAT0_001= 'system=image'
WAT1_001= 'wtype=tan axtype=ra'
WAT2_001= 'wtype=tan axtype=dec'
LTV1 = -24.
LTV2 = -24.[/b:87d4a23657]and ccxymatch DOES NOT FIND stars from this header:[b:87d4a23657]EQUINOX = 2000. / Standard FK5 (years)
RADECSYS= 'FK5 ' / Coordinate reference frame
CRVAL1 = 9.24108 / 00:36:57.8, RA at ref pixel
CRVAL2 = -33.54489 / -33:32:41.6, DEC at ref pixel
CRPIX1 = 488. / Ref pixel in X
CRPIX2 = 488. / Ref pixel in Y
CDELT1 = -0.00008000 / SS arcsec per pixel in RA
CDELT2 = 0.00008000 / SS arcsec per pixel in DEC
CTYPE1 = 'RA---TAN' / pixel coordinate system
CTYPE2 = 'DEC--TAN' / pixel coordinate system
PC1_1 = 0.000000 / Translation matrix element
PC1_2 = 1.000000 / Translation matrix element
PC2_1 = -1.000000 / Translation matrix element
PC2_2 = 0.000000 / Translation matrix element
EPOCH = 2000.0
JD = 2453653.75422315
HJD = 2453653.75859356
LJD = 2453653.
WCSDIM = 2
CD1_1 = -8.0000000000000E-5
CD2_2 = 8.00000000000000E-5
LTM1_1 = 1.
LTM2_2 = 1.
WAT0_001= 'system=image'
WAT1_001= 'wtype=tan axtype=ra'
WAT2_001= 'wtype=tan axtype=dec'
LTV1 = -24.
LTV2 = -24.[/b:87d4a23657]The coordinate system in both headers is identical except that they've been dithered. I have also verified that at least some of the RA DEC stars in my list are actually there on each frame I have. So, the only problem is that the triangles algorithm works sometimes and sometimes doesn't. I need it to work all the time, is there something I'm missing, something I can do to ensure that it doesn't fail? Or maybe I should use some other procedure altogether?Thank you in advance for your help and time./aoxsic
lpar of ccxymatch is [b:87d4a23657]
input = SExtractors.coo The input pixel coordinate lists
reference = 2MASS.coo The input celestial coordinate lists
output = db The output matched coordinate lists
tolerance = 1.0 The matching tolerance in arcseconds
ptolerance = 3.0 The matching tolerance in pixels
(refpoints = ) The optional list of reference points
(xin = INDEF) The X coordinate of the reference point
(yin = INDEF) The Y coordinate of the reference point
(xmag = INDEF) The X axis scale in arcseconds / pixel
(ymag = INDEF) The Y axis scale in arcseconds / pixel
(xrotation = INDEF) The X axis rotation in degrees
(yrotation = INDEF) The Y axis rotation in degrees
(projection = tan) The sky projection geometry
(lngref = INDEF) The ra / longitude of the reference point
(latref = INDEF) The dec / latitude of the reference point
(lngcolumn = 1) The reference list ra / longitude coordinate column
(latcolumn = 2) The reference list dec / latitude coordinate column
(xcolumn = 1) The pixel list x coordinate column
(ycolumn = 2) The pixel list y coordinate column
(lngunits = degrees) The ra / longitude units
(latunits = degrees) The dec / latitude units
(separation = 3.0) The minimum object separation in arcseconds
(pseparation = 9.0) The minimum object separation in pixels
(matching = triangles) The matching algorithm
(nmatch = 90) The maximum number of points for triangles algorithm
(ratio = 10.0) The maximum ratio of longest to shortest side of triangle
(nreject = 10) The maximum number of rejection iterations
(lngformat = ) The format of the output ra / longitude coordinate
(latformat = ) The format of the output dec / latitude coordinate
(xformat = %13.3f) The format of the output x coordinate
(yformat = %13.3f) The format of the output y coordinate
(verbose = yes) Verbose mode ?
(mode = ql)
[/b:87d4a23657]

It has been my experience that ccxymatch is difficult to find parameters that work in all cases. In other words, you experience of finding it works in one case but not in a fairly similar second case is not unusual. It would be nice to work on making this task more robust but we don't have the resources for this. There are ways to make this work better.You might look at msccmatch in the mscred package. This is less general in that images are expected to be offset only, with some low order geometric difference due to refraction or filter scale changes. While it was written for mosaics it should work well with single images too.Yours,
Frank Valdes