iraf.net

I am reducing data from the Blanco + Mosaic2. The pipeline data has a pretty good WCS, but I am trying to improve it for critical matching, using a list of astrometric stars from the UCAC4, all with position uncertainties < 100 mas.I can do good solutions using msctpeak, but what should be the easier route, using msccmatch, fails (see below).I'm running IRAF 2.14 (for reasons I won't go into now) on a Mac, Sys 10.6.8Thanks!
Frank W.
***************************************************************
Here is part of header from a typical image (chip 1):b_tu284827[1][2048,4096][real]: SN1006_n4_p13
No bad pixels, min=0., max=0. (old)
Line storage mode, physdim [2048,4096], length of user area 14459 s.u.
Created Tue 21:32:11 20-Apr-2010, Last modified Tue 21:32:10 20-Apr-2010
Pixel file "tu284827.fits" [ok]
RAWFILE = 'obj4051 ' / Original raw file
FILENAME= 'C4M10A_20100414-B-ct4m20100418T072247' / Current filename
NEXTEND = 8 / Number of extensions
OBSTYPE = 'object ' / Observation type
PROCTYPE= 'InstCal ' / Processing type
PRODTYPE= 'image ' / Product type
MIMETYPE= 'application/fits' / Mimetype of this data file
RADECSYS= 'FK5 ' / Default coordinate system
RA = '15:02:48.72' / [h] RA center of exposure
DEC = '-41:56:17.0' / [deg] DEC center of exposure
CENTDEC = -41.93805 / [deg] DEC center of exposure
CORN1RA = 225.2882 / [deg] RA corner of exposure
CORN2RA = 225.3043 / [deg] RA corner of exposure
CORN3RA = 226.1178 / [deg] RA corner of exposure
CORN4RA = 226.0942 / [deg] RA corner of exposure
CENTRA = 225.703 / [deg] RA center of exposure
CORN1DEC= -41.6396 / [deg] DEC corner of exposure
CORN2DEC= -42.2512 / [deg] DEC corner of exposure
CORN3DEC= -42.2355 / [deg] DEC corner of exposure
CORN4DEC= -41.6249 / [deg] DEC corner of exposureTIMESYS = 'UTC ' / Time system
DATE-OBS= '2010-04-18T07:22:47.3' / Date and time of exposure start
TIME-OBS= '07:22:47.329' / Time of observation start
MJD-OBS = 55304.30749223 / MJD of observation start
MJDHDR = 55304.30742824 / MJD of header creation
LSTHDR = '16:24:41.8 ' / LST of header creationOBSERVAT= 'CTIO ' / Observatory
ADC = 'Blanco ADC' / ADC IdentificationINSTRUME= 'mosaic_2' / Detector
FILPOS = 3 / Filter position
FILTER = 'B Harris c6002 ' / Filter name(s)
FILTID = 'c6002 ' / Unique filter identificationOBSERVER= 'Winkler & Long' / Observer(s)
PROPOSER= 'P. Frank Winkler' / Proposer(s)
PROPOSAL= 'Young SNRs, Nonradiative Shocks, and Cosmic Rays: Measuring the Expa'
PROPID = '2010A-0483' / Proposal identification
OBSID = 'ct4m.20100418T072247' / Observation IDDTSITE = 'ct ' / observatory location
DTTELESC= 'ct4m ' / telescope identifier
DTINSTRU= 'mosaic_2 ' / instrument identifier
DTCALDAT= '2010-04-17 ' / calendar date from observing schedule
ODATEOBS= ' ' / previous DATE-OBS
DTUTC = '2010-04-18T07:28:42' / post exposure UTC epoch from DTS
DTOBSERV= 'NOAO ' / scheduling institution
DTPROPID= '2010A-0483 ' / observing proposal ID
DTPI = 'P. Frank Winkler ' / Principal Investigator
DTPIAFFL= 'Middlebury College' / PI affiliation
DTTITLE = 'Young SNRs, Nonradiative Shocks, and Cosmic Rays: Measuring the Expa'
DTCOPYRI= 'AURA ' / copyright holder of data
DTACQUIS= 'ctioa8.ctio.noao.edu' / host name of data acquisition computer
DTACCOUN= 'lp ' / observing account name
DTACQNAM= '/ua84/mosaic/winkler/n4/obj4051.fits' / file name supplied at teles
DTNSANAM= 'ct2488452.fits ' / file name in NOAO Science Archive
DTSTATUS= 'done ' / data transport status
SB_HOST = 'pipedmn.tuc.noao.edu' / iSTB client host
SB_ACCOU= 'pipeline ' / iSTB client user account
SB_SITE = 'tu ' / iSTB host site
SB_LOCAL= 'tu ' / locale of iSTB daemon
SB_DIR1 = 'Q20100420 ' / level 1 directory in NSA DS
SB_DIR2 = 'C4M10A ' / level 2 directory in NSA DS
SB_DIR3 = '20100414 ' / level 3 directory in NSA DS
SB_RECNO= 284827 / iSTB sequence number
SB_ID = 'tu284827 ' / unique iSTB identifier
SB_NAME = 'tu284827.fits ' / name assigned by iSTB
RMCOUNT = 1 / remediation counter
PHOTFILT= 'g '
MAGZREF = 25.41
MAGZFLT = 'B Harris c6002'
EXPTIME = 200.000 / Exposure time in secsWAT1_001= 'wtype=zpx axtype=ra projp0=0. projp1=1. projp2=0. projp3=337.74 proj'
WAT2_001= 'wtype=zpx axtype=dec projp0=0. projp1=1. projp2=0. projp3=337.74 pro'PIPELINE= 'NOAO Mosaic Pipeline' / Name of calibration pipeline
PLVER = 'MOSAIC V1.1' / Pipeline version
ASTRMCAT= 'USNO-B1 ' / Catalog used for astrometric calibration
EFFTIME = 1. / Effective exposure time
WCSAXES = 2 / Number of WCS axes
PLPROPID= '2010A-0483'PLQUEUE = 'C4M10A ' / PL Queue
PLQNAME = '20100414' / PL Dataset
PLPROCID= '7949826 ' / PL Processing ID
PLFNAME = 'B-ct4m20100418T072247' / PL Filename
PLOFNAME= 'obj4051_pl' / Original file name
PCOUNT = 0 / No 'random' parameters
GCOUNT = 1 / Only one group
ORIGIN = 'NOAO-IRAF FITS Image Kernel July 2003' / FITS file originator
EXTNAME = 'ccd1 ' / Extension name
INHERIT = T / Inherits global header
OBJECT = 'SN1006_n4_p13' / Name of the object observed
DATE = '2010-04-21T02:32:11' / Date FITS file was generated
IRAF-TLM= '2010-04-21T02:32:10' / Time of last modification
IMAGEID = 1 / Image identificationEQUINOX = 2000. / Equinox of WCS
CTYPE1 = 'RA---ZPX' / Coordinate type
CTYPE2 = 'DEC--ZPX' / Coordinate type
CRVAL1 = 225.70170833334 / Coordinate reference value
CRVAL2 = -41.942861111111 / Coordinate reference value
CRPIX1 = 4116.21494333764 / Coordinate reference pixel
CRPIX2 = 4132.5708027145 / Coordinate reference pixel
CD1_1 = -2.0853182355153E-6 / Coordinate matrix
CD2_1 = 7.47002348777037E-5 / Coordinate matrix
CD1_2 = 7.47832189052448E-5 / Coordinate matrix
CD2_2 = 1.94745408581496E-6 / Coordinate matrix
WAT0_001= 'system=image' / Coordinate system
WAT1_002= 'p4=0. projp5=632052. lngcor = "3. 3. 3. 2. -0.3046592774256093 0.005'
WAT1_003= '846650419468163 -0.3154535672801381 -0.1545673620040663 1.2961220949'
WAT1_004= '18725E-4 0.002485071865706102 0.00389236219528579 -2.202923767390268'
WAT1_005= 'E-4 0.005625335647979927 -0.002256791838819684 "'
WAT2_002= 'jp4=0. projp5=632052. latcor = "3. 3. 3. 2. -0.3046592774256093 0.00'
WAT2_003= '5846650419468163 -0.3154535672801381 -0.1545673620040663 4.640687600'
WAT2_004= '194133E-4 0.001146416910324128 0.003053684921863262 0.00361996704922'
WAT2_005= '0075 9.954653068438310E-4 0.007385201317320082 "'RA1 = '15:02:00.41' / Mosaic CCD center
DEC1 = '-42:10:32.9' / Mosaic CCD center
CENRA1 = 225.5017 / Mosaic CCD center
COR1RA1 = 225.3043 / Mosaic CCD corner
COR2RA1 = 225.2962 / Mosaic CCD corner
COR3RA1 = 225.704 / Mosaic CCD corner
COR4RA1 = 225.7095 / Mosaic CCD corner
CENDEC1 = -42.1758 / Mosaic CCD center
COR1DEC1= -42.2512 / Mosaic CCD corner
COR2DEC1= -42.1026 / Mosaic CCD corner
COR3DEC1= -42.097 / Mosaic CCD corner
COR4DEC1= -42.2472 / Mosaic CCD cornerCHECKSUM= 'Q4CFS49EQ4CEQ49E' / HDU checksum updated 2012-12-31T22:43:04
DATASUM = '701863915' / data unit checksum updated 2012-12-31T22:43:04
HISTORY Image was compressed by CFITSIO using scaled integer quantization:
HISTORY q = 4.000000 / quantized level scaling parameter
HISTORY 'SUBTRACTIVE_DITHER_1' / Pixel Quantization Algorithm********** end of header info *************************************
Start of coordinate file: sn1006_ucac4_edit.txt
15.05729533 -41.7804237 12.504
15.05820049 -42.319262 12.505
15.01622726 -42.1308992 12.507
15.01081683 -41.9395145 12.517
15.08093916 -42.3082123 12.527
15.07272389 -41.8033912 12.529
15.07518245 -42.0264153 12.533
15.06811861 -42.2258873 12.551
15.01105845 -42.3323564 12.566
15.09074999 -42.0905639 12.567
15.03330843 -41.5982689 12.571
15.05404791 -41.5194445 12.58
...
(the file has about 1600 stars, covering all dither positions for all chips)I run msccmatch, with the following parameters:
MSCCMATCH params:
input = "b_tu284827" List of input mosaic exposures
coords = "sn1006_ucac4_edit.txt" Coordinate file (ra/dec)
accept = yes Accept solution?
(outcoords = "") List of updated coordinate files
(usebpm = yes) Use bad pixel masks?
(verbose = yes) Verbose?\n\n# Coarse Search
(nsearch = 0) Maximum number of positions to use in search
(search = 0.) Translation search radius (arcsec)
(rsearch = 0.) Rotation search radius (deg)\n\n# Fine Centroid
(cbox = 15) Centering box (pixels)
(maxshift = 1.) Maximum centering shift to accept (arcsec)
(csig = 0.1) Maximum centering uncertainty to accept (arcsec
(cfrac = 0.03) Minimum fraction of accepted centers
(listcoords = yes) List centered coordinates in verbose mode?\n\n#
(nfit = 20) Min for fit (>0) or max not found (<=0)
(rms = 1.) Maximum fit RMS to accept (arcsec)
(fitgeometry = "general") Fitting geometry
(reject = 3.) Fitting rejection limit (sigma)
(update = yes) Update coordinate systems?
(interactive = yes) Interactive?
(fit = yes) Interactive fitting?
(graphics = "stdgraph") Graphics device
(cursor = "") Graphics cursor\n
(mode = "q") Output (last of many lines) from msccmatch:
b_tu284827[8] -6.2839E8 -1.00 0.00 -1.81E8 8.00 0.00
b_tu284827[8] -2.2729E9 0.02 6.00 0.00 0.00 INDEF **101**
b_tu284827[8] -2.6320E9 0.09 2.00 0.00 0.00 INDEF **101**
b_tu284827[8] 16308839.0 -13.00 -1.39E9 0.00 0.00 0.00 **INDEF**
b_tu284827[8] 2.56841E8 -13.00 -1.31E9 0.00 0.00 0.00 **INDEF**
b_tu284827[8] -3.7753E7 10355258.0 0.00 0.00 INDEF INDEF **OffImage**
b_tu284827[8] -5.2648E8 -9.00 0.00 -9.71E8 0.00 0.00
b_tu284827[8] -1.6502E8 -2.00 0.00 -2.70E6 50.0 0.00 **INDEF**
b_tu284827[8] 1.95976E9 -7.00 0.00 -7.28E8 6.00 0.00
b_tu284827[8] 2.54917E9 1.07376E8 0.00 0.00 INDEF INDEF **OffImage**
b_tu284827[8] 1.45491E8 1.21027E9 0.00 0.00 INDEF INDEF **OffImage**
b_tu284827[8] 0.00 -4.0490E8 0.00 0.00 0.00 INDEF **101**
b_tu284827[8] 3.01006E9 -1.00 0.00 -2.E100 0.00 0.00 **INDEF**
b_tu284827[8] 0.00 2.07583E9 4.00 0.00 0.00 INDEF **101**
b_tu284827[8] 0.00 1.91502E9 6.00 0.00 0.00 INDEF **101**
401/582 not centroided
Warning: Too many coordinates failed to centroid: 401/582 < 0.03
ERROR: MSCCMATCH failed for b_tu284827

Hi Frank,I'm sorry for the delay but I had a high priority project last week. First concerning zpx and msctpeak. Yes this is the preferred WCS for MOSAIC data but TNX can also be used. To create a zpx WCS with msctpeak the projection parameter is set to a file containing projection coefficients. This is often not obvious. The most recent values used for CTIO are [code:1:130b2fe3fa]
zpx
projp0 0.
projp1 1.
projp2 0.
projp3 337.74
projp4 0.
projp5 632052.
[/code:1:130b2fe3fa]So you would copy this out to something like "zpx.dat" and specify this file in the parameter. You can also find the coefficients in the WAT keywords.The ultimate cause of the msccmatch failure seems to be too many centroid failures. I just noticed the inequality in the error report is in the wrong direction. With a value of cfrac of 0.03 it should say 401/582 > 0.03 which means more than 17 sources failed to centroid. The value 0.03 is rather small.What I also notice is the second and third columns of the individual centroiding report output don't look like x/y pixel values as they should. I am going to make the guess that a bug in the zpx inversion code (going from RA/DEC -> x/y) which was fixed in V2.15 is causing the conversion from the input RA/DEC in your coordinate file to be inverted grossly incorrectly. The only work around (other than updating IRAF) I can think of is to convert the WCS to tnx but this can be a little complicated and you would need to check that the coordinate evaluations are working correctly.What you would need to do is make a grid of x/y values over the CCDs. Duplicate the columns so that you have "x y x y". Then use wcsctran to convert two columns to RA/DEC based on the calibrated ZPX system. Then you would use that as input to ccmap to make a new WCS with TNX as the projection. You would then apply the database solution (or just let ccmap update the image WCS).So that is my interpretation of the problem.Frank

Hi Frank V-Don't beat yourself up about the delay; not to worry.I have tried the experiment of using msctpeak with projection = zpx.dat
(the file you provided), and this DOES work. Happily, the results are virtually
indistinguishable from what I found using a tnx projection (both with all
orders set = 4, cross terms to half). One detail question though: does it matter whether the fitting function is
polynomial, legendre, or chebyshev? (I used polynomial in both cases, I believe).As for msccmatch: I will soon have a machine upgraded to IRAF 2.15, so the
bug you mention should be fixed, and I'll proceed then. I don't want to go to
ver 2.16 because apparently this is not compatible with Gemini software.Thanks very much,
Frank W.

In practice I don't think the choice of function makes in difference. The others are more "orthogonal" which has implications for robust fitting but if a fit is made they should all evaluate equally well.The one thing about orders is that sometimes the function is evaluated outside the image (such as determining when an RA/DEC is in the image). If the order is to high the extrapolation outside where the fit was done can go wonky (which is basically the problem with zpx bug). So you should not go beyond order 4. The advantage of zpx is that the radial distortion is model mostly away and so the fitting of the residuals by a polynomial is less extreme.Frank

Thanks--
That's pretty much what I figured. I never go higher than order 4, for exactly the reasons you mention. I'll try using zpx in future, now that I know the key to doing it.Frank W