False positive

[paul.l...@gmail.com]

I appear to have found a genuine case of a false positive identification in that a single FITS file will produce radically different WCS values according to the level of down sampling with the -z option to solve-field. I would attach the file but for its size:

pcl@horus ~/astro/M33 $ ls -lsh foo.fits
140M -rw-r--r-- 1 pcl pcl 140M Feb 11 10:07 foo.fits
pcl@horus ~/astro/M33 $ identify foo.fits
foo.fits[0] FITS 6032x4032 6032x4032+0+0 16-bit Grayscale Gray 145.9MB 1.540u 0:01.540
foo.fits[1] FITS 6032x4032 6032x4032+0+0 16-bit Grayscale Gray 145.9MB 1.050u 0:01.050
foo.fits[2] FITS 6032x4032 6032x4032+0+0 16-bit Grayscale Gray 145.9MB 0.550u 0:00.550

Here's the report for -z 8

pcl@horus ~/astro/M33 $ time solve-field -z 8 -K -p foo.fits &
[2] 15628
pcl@horus ~/astro/M33 $ Reading input file 1 of 1: "foo.fits"...
Extracting sources...
This looks like a multi-color image: processing the first image plane only. (NAXIS=3)
Downsampling by 8...
simplexy: found 509 sources.
Solving...
Reading file "./foo.axy"...
log-odds ratio 156.949 (1.45231e+68), 57 match, 0 conflict, 91 distractors, 80 index.
RA,Dec = (22.8644,30.4697), pixel scale 3.06535 arcsec/pix.
Hit/miss: Hit/miss: ++-+++++++++++++++++-++++++-++-++++--+-+-+-+-+-+++--------+----++-+--++--+-+--++----------+---------
Field 1: solved with index index-4211.fits.
Field 1 solved: writing to file ./foo.solved to indicate this.
Field: foo.fits
Field center: (RA,Dec) = (22.863853, 30.476194) deg.
Field center: (RA H:M:S, Dec D:M:S) = (01:31:27.325, +30:28:34.299).
Field size: 5.15412 x 3.44867 degrees
Field rotation angle: up is 86.1826 degrees E of N
Field parity: pos
Creating new FITS file "./foo.new"...


real 0m12.689s
user 0m11.883s
sys 0m0.487s

and the same for no down-sampling:

pcl@horus ~/astro/M33 $ time solve-field -K -p foo.fits &
[2] 15796
pcl@horus ~/astro/M33 $ Reading input file 1 of 1: "foo.fits"...
Extracting sources...
This looks like a multi-color image: processing the first image plane only. (NAXIS=3)
simplexy: found 219 sources.
Solving...
Reading file "./foo.axy"...
log-odds ratio 25.6102 (1.32551e+11), 14 match, 5 conflict, 38 distractors, 26 index.
RA,Dec = (214.311,19.135), pixel scale 0.07219 arcsec/pix.
Hit/miss: Hit/miss: +-+---+-+---c-----++--+-c-------+-+-c---+-+-c---c---++--+(best)-----------------c---------c---c-----------
fit-wcs.c:118:fit_sip_wcs: Too few correspondences for the SIP order specified (4 < 6)

Field 1: solved with index index-4202-08.fits.
log-odds ratio 25.6102 (1.32551e+11), 14 match, 5 conflict, 38 distractors, 26 index.
RA,Dec = (214.311,19.135), pixel scale 0.07219 arcsec/pix.
Hit/miss: Hit/miss: +-+---+-+---c-----++--+-c-------+-+-c---+-+-c---c---++--+(best)-----------------c---------c---c-----------
fit-wcs.c:118:fit_sip_wcs: Too few correspondences for the SIP order specified (4 < 6)

Field 1: solved with index index-4202-08.fits.
Field 1 solved: writing to file ./foo.solved to indicate this.
Field: foo.fits
Field center: (RA,Dec) = (214.311338, 19.134970) deg.
Field center: (RA H:M:S, Dec D:M:S) = (14:17:14.721, +19:08:05.890).
Field size: 7.2575 x 4.85116 arcminutes
Field rotation angle: up is 2.27658 degrees E of N
Field parity: neg
Creating new FITS file "./foo.new"...


real 0m35.214s
user 0m34.223s
sys 0m0.967s

The first is correct.  The FITD file was created from a NEF-format raw image of the M33 region in Triangulum.  The second WCS is way out in the weeds.

Any comments and/or suggestions?

Paul

[Dustin Lang]

Have a look at the source detection images -- if you have the plotting programs, the *-objs.png plot will show you the detected sources.  I would guess that there is a line of false sources in your image that matches a false line of sources in the reference catalog.  That's a typical failure mode.

cheers,
--dustin