Debian build does not find any NGC / Messier objects

158 views
Skip to first unread message

Kaori Takahashi-Riekko

unread,
Dec 2, 2016, 12:16:27 PM12/2/16
to astrometry
Hello,


first, warm thanks for the creators of the software and the maker(s) of the Debian package.

The installation was easy and I got solve-field up and running in minutes.
It correctly locates the photos I've taken and marks the prominent stars on it, but for some reason it never marks any NGC / Messier objects on them, even though the online service finds and marks them correctly.

The command I used was:
solve-field -u degwidth -L 10 -H 14 --parity neg input.jpg

I was wondering if I'm missing a catalog or something due to Debian's strict copyright policies, but couldn't find anything relevant in Google.
Next I tried plot-constellations with the following command:

plot-constellations -N -C -B -v -w input.wcs -o solved.png

I got the following output:
Trying to parse SIP/TAN header from input.wcs...
Got SIP header.
Checking 88 constellations.
done constellations.
Checking 5023 bright stars.
14Aur at (4903.23, 3295.89)
The star 14Aur
16Aur at (4542.73, 2918.85)
The star 16Aur
19Aur at (4302.61, 2591.21)
The star 19Aur
σAur at (3654.43, 639.899)
The star σAur
φAur at (3376.63, 2328)
The star φAur
χAur at (2797.1, 3664.22)
The star χAur
υAur at (619.653, 695.765)
The star υAur
Checking 0 NGC/IC objects.

The online service recognises and marks for example M36, M38 and NGC1907 on the same photo.
I have to admit that I don't have a clue how to proceed further.
solve-field is version 0.67 and apt-get update doesn't list any newer versions.

I'd be really happy if anyone could help me ^.^
Clear skies,


Kaori

Dustin Lang

unread,
Dec 2, 2016, 12:23:57 PM12/2/16
to astrometry
Hi,

Could you please send a link to the web version of this image?  Just helpful for testing.

The "Checking 0 NGC/IC objects" is suspicious... that catalog is supposed to be compiled in...

thanks,
--dustin


Dustin Lang

unread,
Dec 2, 2016, 12:28:09 PM12/2/16
to astrometry
Ahhh, I see in my notes from Ole Streicher, the Debian maintainer,

" I still have to remove the ngc2000 catalog because of the "non-commercial" clause -- I replace it with an empty catalog (/dev/null), so this will just remove the names from the final image. However, I also started a request to the copyright holder (which is F+W Media today) to make it free -- no response so far, I'll probably need to send a reminder."

so that explains why the Debian version lacks the NGC2000 labels...

cheers,
--dstn

Kaori Takahashi-Riekko

unread,
Dec 2, 2016, 1:13:15 PM12/2/16
to astrometry
Hi Dustin,

came here to post the links to the online submissions, but it seems that you've already found the cause =)
It seems that the best course of action is to compile it myself from source?

Thank you for your help!


--Kaori

Dustin Lang

unread,
Dec 2, 2016, 1:26:07 PM12/2/16
to astrometry
I guess so :)


Ole Streicher

unread,
Dec 2, 2016, 2:51:00 PM12/2/16
to astro...@googlegroups.com
To give the current status here: I had a lengthy discussion (which took
~9 months in total) with the F+W people, and they decided finally that
they will not put these names under a free license.

Fortunately, there is now an attempt to create an open NGC2000
catalogue:

https://github.com/mattiaverga/OpenNGC

This looks very promising and will hopefully solve all these ugly
problems. Many thanks to Mattia Verga for this project!

Cheers

Ole

Kaori Takahashi-Riekko

unread,
Dec 2, 2016, 4:26:17 PM12/2/16
to astrometry
OK, managed to compile it myself. It now recognises about third of the NGCs and I think that with some tinkering I'll get it to mark everything for me (✿◠‿◠)
If somebody decides to follow my footsteps before the OpenNGC project is finished and hopefully integrated, here are my notes:

1. Running Debian Sid, clone the GitHub repo (double bleeding edge warning)
2. Dependencies part 1: apt-get install libcairo2-dev libnetpbm10-dev libpng12-dev libjpeg-dev python-numpy python-pyfits python-dev zlib1g-dev libbz2-dev swig
3. Dependencies part 2: download and compile cfitsio from http://heasarc.gsfc.nasa.gov/fitsio/
4. make CFITS_INC="-I/home/kaori/cfitsio/include" CFITS_LIB="-L/home/kaori/cfitsio/lib -lcfitsio" all
5. make CFITS_INC="-I/home/kaori/cfitsio/include" CFITS_LIB="-L/home/kaori/cfitsio/lib -lcfitsio" install INSTALL_DIR=/home/kaori/astrometry3 (yea, needed 3 tries ;_;)
6. Edit /home/kaori/astrometry3/etc/astrometry.cfg
7. Copy index files to the correct directory
8. export PYTHONPATH=/home/kaori/astrometry3/lib/python/
9. Run /home/kaori/astrometry3/bin/solve-field with appropriate parameters
10. Either get results in two minutes or nothing in two hours on a 4.8GHz Sandy. Already ~70% success rate and I'm a complete beginner with this =)

Thank you Dustin for helping me with this and Ole for your 9-month campaign.


--Kaori

Steve Cookson

unread,
Dec 10, 2018, 2:40:48 PM12/10/18
to astrometry
Is there any news on this?  I have Ubuntu Astronomy which is 16.04 and solve-field still has the ngc problem.

It looks like there is now an open catalogue.

What do I need to do to make it spring into life?

Thanks

Steve.

Dustin Lang

unread,
Dec 10, 2018, 3:26:21 PM12/10/18
to astrometry
Someone was working on this way back.  Haven't heard anything lately...



Reply all
Reply to author
Forward
0 new messages