imtranspose error
Brett Gladman wrote on Jan 25, 2010
This error occurs on
PC-IRAF Rev 2.13-beta 2
running on 64-bit Ubuntu 9.10
REPRODUCIBLE ERROR
The command :
imtranspose inputimage[*,-*] outputimage
creates a nonsense output image (small mean and huge standard deviation).
Partial solution:
If the inputimage is transposed first (with no [*,-*]) and then
imcopied with [-*,*] this does not happen.
This workaround is very unattractive (requiring twice as much CPU
time and a temp copy to be created).
Is this a 64-bit problem?
PC-IRAF Rev 2.13-beta 2
running on 64-bit Ubuntu 9.10
REPRODUCIBLE ERROR
The command :
imtranspose inputimage[*,-*] outputimage
creates a nonsense output image (small mean and huge standard deviation).
Partial solution:
If the inputimage is transposed first (with no [*,-*]) and then
imcopied with [-*,*] this does not happen.
This workaround is very unattractive (requiring twice as much CPU
time and a temp copy to be created).
Is this a 64-bit problem?
Mike Fitzpatrick wrote on Jan 25, 2010
Could you post a complete header from the input image? I thought earlier I saw where a "wcsreset physical" would also fix things, but this doesn't sound like a 64-bit problem.
-Mike
-Mike
Brett Gladman wrote on Jan 25, 2010
Hello Mike
sorry for the long delay here. The images in question are Palomar LFC
moscaic camera images. The headers are quite sparse. I have just
re-confirmed that the transpose and then imcopy[*,-*] still works, but
that imtranspose[*,-*] gives a nonsense image.
- Brett
e[1022,2045][real]: SandT
No bad pixels, min=0., max=0. (old)
Line storage mode, physdim [1022,2045], length of user area 2673 s.u.
Created Fri 14:02:12 19-Mar-2010, Last modified Fri 14:02:16 19-Mar-2010
Pixel file "e.fits" [ok]
EXTEND = F / File may contain extensions
ORIGIN = 'NOAO-IRAF FITS Image Kernel July 2003' / FITS file originator
IRAF-TLM= '2010-03-19T21:02:16' / Time of last modification
DATE = '2010-03-19T20:54:53' / UT Date of file creation (DD/MM/YY)
LATITUDE= 33.4000 / Latitude (degrees N)
LONGITUD= -116.9000 / Longitude (degrees E)
TELESCOP= 'Hale 5m Telescope' / Telescope used for observation
FRATIO = 7.5000 / Focal ratio of telescope
INSTRUME= 'LFC ' / Instrument used for observation
DETECTOR= 'SITe SI002 x 6' / Detector used for observation
FRAME = 240 / Frame number of observation
CCDPICNO= 240 / Frame number of observation
OBJECT = 'SandT ' / Name of object
IMAGETYP= 'OBJECT ' / Type of observation
EXPTIME = 360.012 / Integration time (seconds)
DARKTIME= 360.695 / Dark current time (seconds)
DATE-OBS= '2010-03-19' / UT Date of observation (DD/MM/YY)
UT = ' 06:58:22.00' / Universal time (UTC) at exposure start
JD = 2455274.790532 / Julian date at exposure start
RA = ' 12:08:27.01' / Right Ascension
DEC = ' 01:10:59.90' / Declination
EQUINOX = 2000.000 / Equinox of RA and DEC
EPOCH = 2000.000 / Equinox of RA and DEC
HA = '-01:07:27.60' / Hour angle at start
ST = ' 10:57:38.52' / Sidereal time at start
AIRMASS = 1.234 / Airmass at start
FILTER = 'r' ' / Filter description
GAIN = 1.100 / Nominal gain (e-/ADU)
SECPIX1 = 0.363 / Arcseconds per pixel in fast dir
SECPIX2 = 0.363 / Arcseconds per pixel in slow dir
CCDBIN1 = 2 / On-chip column binning (fast dir)
CCDBIN2 = 2 / On-chip row binning (slow dir)
ROTANGLE= 0.00 / Rotation angle (deg)
LOGINFO = ' Status=0x00000000'
CHIPID = 0
ORIENT = 'ew '
WCSDIM = 2
LTM1_1 = 1.
LTM2_2 = 1.
WAT0_001= 'system=physical'
WAT1_001= 'wtype=linear'
WAT2_001= 'wtype=linear'
TRIM = 'Mar 19 3:44 Trim data section is [2:1023,2:2046]'
OVERSCAN= 'Mar 19 3:45 Overscan section is [1030:1054,2:2046] with mean=829.39'
ZEROCOR = 'Mar 19 3:45 Zero level correction image is Zero0'
FLATCOR = 'Mar 19 3:45 Flat field image is Domer with scale=10498.76'
CCDSEC = '[2:1023,2:2046]'
CCDMEAN = 9300.89
CCDMEANT= 953437500
CCDPROC = 'Mar 19 3:45 CCD processing done'
sorry for the long delay here. The images in question are Palomar LFC
moscaic camera images. The headers are quite sparse. I have just
re-confirmed that the transpose and then imcopy[*,-*] still works, but
that imtranspose[*,-*] gives a nonsense image.
- Brett
e[1022,2045][real]: SandT
No bad pixels, min=0., max=0. (old)
Line storage mode, physdim [1022,2045], length of user area 2673 s.u.
Created Fri 14:02:12 19-Mar-2010, Last modified Fri 14:02:16 19-Mar-2010
Pixel file "e.fits" [ok]
EXTEND = F / File may contain extensions
ORIGIN = 'NOAO-IRAF FITS Image Kernel July 2003' / FITS file originator
IRAF-TLM= '2010-03-19T21:02:16' / Time of last modification
DATE = '2010-03-19T20:54:53' / UT Date of file creation (DD/MM/YY)
LATITUDE= 33.4000 / Latitude (degrees N)
LONGITUD= -116.9000 / Longitude (degrees E)
TELESCOP= 'Hale 5m Telescope' / Telescope used for observation
FRATIO = 7.5000 / Focal ratio of telescope
INSTRUME= 'LFC ' / Instrument used for observation
DETECTOR= 'SITe SI002 x 6' / Detector used for observation
FRAME = 240 / Frame number of observation
CCDPICNO= 240 / Frame number of observation
OBJECT = 'SandT ' / Name of object
IMAGETYP= 'OBJECT ' / Type of observation
EXPTIME = 360.012 / Integration time (seconds)
DARKTIME= 360.695 / Dark current time (seconds)
DATE-OBS= '2010-03-19' / UT Date of observation (DD/MM/YY)
UT = ' 06:58:22.00' / Universal time (UTC) at exposure start
JD = 2455274.790532 / Julian date at exposure start
RA = ' 12:08:27.01' / Right Ascension
DEC = ' 01:10:59.90' / Declination
EQUINOX = 2000.000 / Equinox of RA and DEC
EPOCH = 2000.000 / Equinox of RA and DEC
HA = '-01:07:27.60' / Hour angle at start
ST = ' 10:57:38.52' / Sidereal time at start
AIRMASS = 1.234 / Airmass at start
FILTER = 'r' ' / Filter description
GAIN = 1.100 / Nominal gain (e-/ADU)
SECPIX1 = 0.363 / Arcseconds per pixel in fast dir
SECPIX2 = 0.363 / Arcseconds per pixel in slow dir
CCDBIN1 = 2 / On-chip column binning (fast dir)
CCDBIN2 = 2 / On-chip row binning (slow dir)
ROTANGLE= 0.00 / Rotation angle (deg)
LOGINFO = ' Status=0x00000000'
CHIPID = 0
ORIENT = 'ew '
WCSDIM = 2
LTM1_1 = 1.
LTM2_2 = 1.
WAT0_001= 'system=physical'
WAT1_001= 'wtype=linear'
WAT2_001= 'wtype=linear'
TRIM = 'Mar 19 3:44 Trim data section is [2:1023,2:2046]'
OVERSCAN= 'Mar 19 3:45 Overscan section is [1030:1054,2:2046] with mean=829.39'
ZEROCOR = 'Mar 19 3:45 Zero level correction image is Zero0'
FLATCOR = 'Mar 19 3:45 Flat field image is Domer with scale=10498.76'
CCDSEC = '[2:1023,2:2046]'
CCDMEAN = 9300.89
CCDMEANT= 953437500
CCDPROC = 'Mar 19 3:45 CCD processing done'
Mike Fitzpatrick wrote on Jan 25, 2010
I can't reproduce this in any way here using artificial data, the transposed image has the same stats as the input image:
So it doesn't matter whether I overwrite the image or create a new one. One thing you might try is to set the IMTRANSPOSE len_blk parameter to 2045 so it processes a complete row.
Otherwise, could you post the problem image to the anonftp at ftp://iraf.noao.edu/pub so I can try it again?
artdata> mkpattern test.fits ncols=1022 nlines=2045
artdata> imtranspose test.fits[*,-*] test2.fits
artdata> imtranspose test.fits[*,-*] test.fits
artdata> imstat test*
# IMAGE NPIX MEAN STDDEV MIN MAX
test.fits 2089990 96.77 55.85 0. 129.
test2.fits 2089990 96.77 55.85 0. 129.
So it doesn't matter whether I overwrite the image or create a new one. One thing you might try is to set the IMTRANSPOSE len_blk parameter to 2045 so it processes a complete row.
Otherwise, could you post the problem image to the anonftp at ftp://iraf.noao.edu/pub so I can try it again?
Last post on Jan 25, 2010