Local solve fails

[Arto Oksanen]

Hello!

I am using local Astrometry to solve images taken with a wide angle lens and a ccd camera. I am cropping just the center 20x20 degrees of the green channel and saving it as a fits file. The stars are located with findstars.stars() and then solve() is called with the star list. This is working fine with 90 percent of the images, but sometimes there is no solution though the image and star list looks good for me.

 id xcentroid ycentroid sharpness roundness1 roundness2 npix sky  peak flux  mag
--- --------- --------- --------- ---------- ---------- ---- --- ----- ---- -----
  1     166.0      92.8      0.29     -0.237      0.185   25 0.0 255.0  1.8 -0.62
  2      84.7      51.1      0.63     -0.785      0.215   25 0.0 255.0  1.7 -0.59
  3     283.3     296.6      0.71     -0.790     -0.006   25 0.0 255.0  1.5 -0.44
  4     435.2     325.8      0.30     -0.176      0.078   25 0.0 234.0  1.4 -0.36
  5      86.4     375.9      0.71     -0.488      0.070   25 0.0 255.0  1.4 -0.35
  6      79.3     308.3      0.29      0.398      0.177   25 0.0 241.0  1.4 -0.34
  7     340.6     345.2      0.65     -0.734      0.041   25 0.0 255.0  1.3 -0.31
  8     409.9     145.7      0.76      0.065      0.020   25 0.0 210.0  1.3 -0.29
  9     437.1     270.1      0.71     -0.201      0.274   25 0.0 236.0  1.3 -0.28
 10     301.9     230.2      0.76     -0.666      0.009   25 0.0 207.0  1.2 -0.16
 11     481.6      14.5      0.76     -0.931     -0.042   25 0.0 195.0  1.1 -0.15
 12       1.3     141.0      0.62      0.554      0.880   25 0.0 155.0  1.1 -0.12
 13      31.1     187.3      0.79     -0.195      0.090   25 0.0 195.0  1.1 -0.08
 14     380.3     356.6      0.29     -0.413      0.165   25 0.0 255.0  1.0 -0.02
found  14  stars from the image


2026-10-01 08:18:06,367 [INFO] root: loaded 3 index files
2026-10-01 08:18:06,367 [INFO] root: solve 1: start
2026-10-01 08:18:06,367 [INFO] root: solve 1: slice=[0, 14[ (1 / 1), index="4100/index-4116.fits" (1 / 3)
2026-10-01 08:18:06,368 [INFO] root: solve 1: slice=[0, 14[ (1 / 1), index="4100/index-4117.fits" (2 / 3)
2026-10-01 08:18:06,369 [INFO] root: solve 1: slice=[0, 14[ (1 / 1), index="4100/index-4115.fits" (3 / 3)
Failure!

The same star list is solving just fine online: https://nova.astrometry.net/status/16151449

What could be wrong? I am using the same index files (4115, 4116, 4117) ) that the online version is solving successfully. It looks from the log that it is not even trying.

Python code:

  solution_params = astrometry.SolutionParameters()

  solution_params.log_odds_keep = 14.0
  solution_params.sip_order = 2

  solver = astrometry.Solver(
    astrometry.series_4100.index_files(
        cache_directory="astrometry_cache",
        scales={15,16,17}  # Wide-field index scales
    )
)
  solution = solver.solve(stars=stars,
       size_hint=astrometry.SizeHint(
            lower_arcsec_per_pixel=155.0,
            upper_arcsec_per_pixel=165.0,
        ),
        position_hint=astrometry.PositionHint(
            ra_deg=235.0,
            dec_deg=28.0,
            radius_deg=10.0,
        ),
        solution_parameters=solution_params,
    )

stars: [[165.95099089363435, 92.77354011182763], [84.67468989473778, 51.1330634227132], [283.310186716766, 296.62743485425443], [435.15500723543704, 325.8278853030001], [86.44452723978571, 375.91744479056183], [79.28502588535576, 308.32710519047083], [340.59896344962505, 345.2291862154199], [409.89698052780585, 145.6628851852806], [437.13963732332263, 270.1107866007907], [301.88354948982305, 230.16454980646415], [481.5907698868299, 14.452975717269661], [1.2844308835400529, 140.9657836449299], [31.081055625317305, 187.2586063309943], [380.323289486827, 356.6079590359219]]

arto

--

Arto Oksanen
arto.o...@jklsirius.fi
Muurame, Finland

[Dustin Lang]

Hi,

You have a different RA,Dec center here than on the web site.  Also, no scale range limit on the web site.  You could try expanding the scale range a bit and see if that helps.

cheers,

dustin

--
You received this message because you are subscribed to the Google Groups "astrometry" group.
To unsubscribe from this group and stop receiving emails from it, send an email to astrometry+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/astrometry/CAG2UQ8rJQP9GcDkooQXHnJr89gfhvPqCFgBJcAhyJKE2mPp%2B0Q%40mail.gmail.com.

[Arto Oksanen]

The solved scale is 

 sqrt(det(CD))=161.764 [arcsec]

But I will try a larger scale range. Why would a bigger range help?

The both coordinates are well within the image limits, but I will try again with the exact same coordinates. Would a bigger radius help? 

arto

[Arto Oksanen]

Hello Dustin!

I changed only the PositionHint and radius:

        position_hint=astrometry.PositionHint(
            ra_deg=242.0,
            dec_deg=28.0,
            radius_deg=15.0,
        )

 

and the image is solving now:

2026-10-01 19:52:59,223 [INFO] root: loaded 3 index files
2026-10-01 19:52:59,224 [INFO] root: solve 1: start
2026-10-01 19:52:59,224 [INFO] root: solve 1: slice=[0, 14[ (1 / 1), index="4100/index-4116.fits" (1 / 3)
2026-10-01 19:52:59,557 [INFO] root: solve 1: logodds=33.3826, matches=9, conflicts=0, distractors=2, ra=244.007, dec=29.7216, scale=159.776, index="4100/index-4116.fits"
2026-10-01 19:52:59,557 [INFO] root: solve 1: slice=[0, 14[ (1 / 1), index="4100/index-4117.fits" (2 / 3)
2026-10-01 19:52:59,763 [INFO] root: solve 1: logodds=49.4344, matches=9, conflicts=0, distractors=2, ra=244.052, dec=29.7542, scale=161.4, index="4100/index-4117.fits"
2026-10-01 19:52:59,973 [INFO] root: solve 1: logodds=52.7092, matches=9, conflicts=0, distractors=2, ra=244.071, dec=29.678, scale=162.315, index="4100/index-4117.fits"
2026-10-01 19:53:00,180 [INFO] root: solve 1: logodds=52.5655, matches=9, conflicts=0, distractors=2, ra=244.062, dec=29.7015, scale=161.157, index="4100/index-4117.fits"
2026-10-01 19:53:00,390 [INFO] root: solve 1: logodds=52.3176, matches=9, conflicts=0, distractors=1, ra=243.96, dec=29.753, scale=161.028, index="4100/index-4117.fits"
2026-10-01 19:53:00,390 [INFO] root: solve 1: slice=[0, 14[ (1 / 1), index="4100/index-4115.fits" (3 / 3)
solution.best_match().center_ra_deg=244.07139478721854
solution.best_match().center_dec_deg=29.678014496525527
solution.best_match().scale_arcsec_per_pixel=162.31522706323653
Success!

Thanks!!

arto

[Dustin Lang]

Great!