local solver problems

[Arto Oksanen]

Hello from Finland!

I am trying to solve some SeeStar50 images with the local astrometry code. Some files solve but some do not even though they solve online in one second. I would like to know why.

For example: https://nova.astrometry.net/user_images/15227607#annotated

It fails locally with very little information:

INFO:root:solve 1: start
INFO:root:solve 1: slice=[0, 25[ (1 / 4), index="4100/index-4108.fits" (1 / 2)
INFO:root:solve 1: slice=[0, 25[ (1 / 4), index="4100/index-4107.fits" (2 / 2)
INFO:root:solve 1: slice=[25, 50[ (2 / 4), index="4100/index-4108.fits" (1 / 2)
INFO:root:solve 1: slice=[25, 50[ (2 / 4), index="4100/index-4107.fits" (2 / 2)
INFO:root:solve 1: slice=[50, 75[ (3 / 4), index="4100/index-4108.fits" (1 / 2)
INFO:root:solve 1: slice=[50, 75[ (3 / 4), index="4100/index-4107.fits" (2 / 2)
INFO:root:solve 1: slice=[75, 100[ (4 / 4), index="4100/index-4108.fits" (1 / 2)
INFO:root:solve 1: slice=[75, 100[ (4 / 4), index="4100/index-4107.fits" (2 / 2)
No match found!

The code has these "hints":

  solver = astrometry.Solver(astrometry.series_4100.index_files(
        cache_directory="astrometry_cache",
        scales={7,8}))

  solution = solver.solve(stars=stars,
       size_hint=astrometry.SizeHint(
            lower_arcsec_per_pixel=4.0,
            upper_arcsec_per_pixel=5.0,
        ),
        position_hint=astrometry.PositionHint(
            ra_deg=240.0,
            dec_deg=25.9,
            radius_deg=1.0,
        ),        solution_parameters=astrometry.SolutionParameters(),
    )

best regards,

arto

--

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

[Dustin Lang]

Hi,

In the web version I see:

Field 1: solved with index index-4108.fits.
Pixel scale:	2.38 arcsec/pixel

so it looks like your pixel scale range of 4 to 5 arcsec per pixel is out of that range.

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/CAG2UQ8r-ZD8EWW5jPOL6m%3DjJLytXOjjoBPjHae5P%3D%3D0W7Cq_sA%40mail.gmail.com.

[Dustin Lang]

(replying on-list)...

Hi,

[You're downsampling the images first, so that scale range is right.]

It's easy to add more logging ("-v" on solve-field), but to really answer the question you're asking, why didn't this solve, is not as easy.  The logging will basically tell you "looking for an answer ... chug chug chug ... didn't find an answer".

When you have a successful solve with the web version, then things are much easier.  The web version doesn't exactly run solve-field, but it does run all the same parts.

The major difference from the solve-field default are that the web version does "--downsample 2" by default.

The first things to look at are:

- source extraction.  On the web version, check out the "extraction" tab of the images.  You should see the brightest sources correctly circled with the large bold red circles.  With solve-field, if you have the plotting working, then there should be an "*-objs.png" plot, and that should look the same.

- index files - looks like you've got this

- command-line args - not dropping a correct solution

You sent some code that seems to be in some kind of wrapper script or something.  Does this eventually call solve-field?  If so, can you send the full command that it is running?

cheers,

dustin

[Arto Oksanen]

Hello Dustin!

I am running all my code in python. On the command line solve-field does work with the same image.

oksanen@murtoinen:~/astrometry.net-0.98$ solve-field --overwrite ../tcrb/seestar50/green.fits
Reading input file 1 of 1: "../tcrb/seestar50/green.fits"...
Extracting sources...
simplexy: found 84 sources.
Solving...
Reading file "../tcrb/seestar50/green.axy"...
Field 1 did not solve (index index-4119.fits, field objects 1-10).
Field 1 did not solve (index index-4118.fits, field objects 1-10).
Field 1 did not solve (index index-4117.fits, field objects 1-10).
Field 1 did not solve (index index-4116.fits, field objects 1-10).
Field 1 did not solve (index index-4115.fits, field objects 1-10).
Field 1 did not solve (index index-4114.fits, field objects 1-10).
Field 1 did not solve (index index-4113.fits, field objects 1-10).
Field 1 did not solve (index index-4112.fits, field objects 1-10).
Field 1 did not solve (index index-4111.fits, field objects 1-10).
Field 1 did not solve (index index-4110.fits, field objects 1-10).
Field 1 did not solve (index index-4109.fits, field objects 1-10).
mo field stars:
  star 5; field_xy 95.5,619.9, field_orig 95.5,619.9
  star 4; field_xy 485.7,328.7, field_orig 485.7,328.7
  star 2; field_xy 131.1,351.1, field_orig 131.1,351.1
  star 6; field_xy 419.9,502.3, field_orig 419.9,502.3
  log-odds ratio 179.949 (1.41529e+78), 28 match, 0 conflict, 17 distractors, 32 index.
  RA,Dec = (240.089,25.9778), pixel scale 4.76085 arcsec/pix.
  Hit/miss:   Hit/miss: ++++-++++++++++++--++++-+-+---++--+-----+--++(best)-----------------------------c-----++++

Field 1: solved with index index-4108.fits.

Field 1 solved: writing to file ../tcrb/seestar50/green.solved to indicate this.
Field: ../tcrb/seestar50/green.fits
Field center: (RA,Dec) = (240.087962, 25.977178) deg.
Field center: (RA H:M:S, Dec D:M:S) = (16:00:21.111, +25:58:37.842).
Field size: 42.7641 x 76.0892 arcminutes
Field rotation angle: up is 171.752 degrees E of N
Field parity: neg
Creating new FITS file "../tcrb/seestar50/green.new"...
Creating index object overlay plot...
Creating annotation plot...

It also solves online: https://nova.astrometry.net/user_images/15239972#annotated

But with the python code I dont get any solution for this image though most other images are solving just fine. 

Maybe I should change my pipeline from the python only to a shell script and run the command line solve-field instead.

Btw, the solve-field works only from the installation directory. If I am running it from the image directory I am getting this error:

oksanen@murtoinen:~/tcrb/seestar50$ solve-field --overwrite green.fits
Reading input file 1 of 1: "green.fits"...
/usr/bin/python3: Error while finding module specification for 'astrometry.util.image2pnm' (ModuleNotFoundError: No module named 'astrometry.util')
augment-xylist.c:590:backtick Failed to run command: /usr/bin/python3 -m astrometry.util.image2pnm --infile green.fits --uncompressed-outfile /tmp/tmp.uncompressed.VolGiM --outfile /tmp/tmp.ppm.f0CxBL --ppm --mydir /usr/bin/solve-field
 ioutils.c:567:run_command_get_outputs Command failed: return value 1

arto 

[Dustin Lang]

Hi,

What python code are you using here?  This is not from the astrometry.net codebase, so you must be using some wrapper code from somewhere.  I don't know

And it looks like that wrapper code is not working, if the solve-field command works.

I'm guessing you are using 

https://github.com/neuromorphicsystems/astrometry/

This is not my project, I do not know who this author is, and you should ask your question at that github repo.

cheers,

dustin

[Arto Oksanen]

Yes, that is what I am using (pip install astrometry). Well, it is working, but fails to solve some images. Sorry for writing on wrong mailing list.

br,

arto

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

[Dustin Lang]

If it fails to solve some images that work with solve-field, then I would say it is not working correctly :)

I'm sorry for the confusion - I didn't really know about that project until today.  Good luck getting an answer!

Best wishes,

dustin

[Dustin Lang]

This will depend on how the code is installed.  The Astrometry.net code package's Makefile should install things correctly.  Who knows what this "pip install astrometry" package does.