×

INDI Library v2.0.7 is Released (01 Apr 2024)

Bi-monthly release with minor bug fixes and improvements

Read More...

[SOLVED] Ekos plate solver woes

  • Posts: 1029
  • Thank you received: 301

Replied by Eric on topic Re:Ekos plate solver woes

Would it be possible to have the Veil FITS you used for this log, so that we have the same inputs? I will run the same command on my Linux system for comparison.

-Eric
5 years 6 months ago #30559

Please Log in or Create an account to join the conversation.

  • Posts: 300
  • Thank you received: 57

Replied by Scott Denning on topic Re:Ekos plate solver woes


Sure! Here's a link:
www.dropbox.com/s/ccnqt6n0bm5xuie/Veil_L...-04-32_014.fits?dl=0

I'm going to learn how to use the /usr/bin/solve-field command line and try to debug the problem without Kstars and Ekos. Then maybe I will understand why the solver fails inside of Ekos.

I *really* appreciate your help!
5 years 6 months ago #30562

Please Log in or Create an account to join the conversation.

  • Posts: 460
  • Thank you received: 69

Replied by Jerry Black on topic Re:Ekos plate solver woes


I just figured this out yesterday...
Command line arguments to solve a fits file using solve-field
# had to do this once
sudo easy_install pyfits
 
# export kstars bin path
export PATH=/Applications/kstars.app/Contents/MacOS/netpbm/bin:/Applications/kstars.app/Contents/MacOS/python/bin:/usr/bin:/bin:/usr/sbin:/sbin:/opt/X11/binexport PATH=/Applications/kstars.app/Contents/MacOS/netpbm/bin:/Applications/kstars.app/Contents/MacOS/python/bin:/usr/bin:/bin:/usr/sbin:/sbin:/opt/X11/bin
# export kstars python path
export PYTHONPATH=/Applications/kstars.app/Contents/MacOS/python/bin
 
/Applications/kstars.app/Contents/MacOS/astrometry/bin/solve-field -O  --no-plots —plot-scale 0.25 --no-verify --resort --downsample 2 -L 0.5 -H 3 -u dw -3 00:43:04 -4 41.25 -5 20 --config /Applications/kstars.app/Contents/MacOS/astrometry/bin/astrometry.cfg /Volumes/Photo1/Andromeda_2018-10-14/integration_DBE.fits

The solve-field command line arguments are documented at: manpages.ubuntu.com/manpages/xenial/man1/solve-field.1.html
The following user(s) said Thank You: Scott Denning
Last edit: 5 years 6 months ago by Jerry Black.
5 years 6 months ago #30564

Please Log in or Create an account to join the conversation.

  • Posts: 1029
  • Thank you received: 301

Replied by Eric on topic Re:Ekos plate solver woes

The part I'm interested in are the options -L and -H, and their relation to your FOV. Basically, these affect which catalog/index files will be used. If you increase the interval, solve-field will use more indexes while searching.
Given your imager is 7376x4928, and your FOV is assumed 600'x400', this makes a rough resolution of 4.9 arcsec/pixel. So theory would tell that the options seen in the log would be OK, and that there is even room for improvement in terms of search time.
Now, there is the parity option, which tells the solver if the frame is optically inverted or not compared to the catalog data. Normally the first solve is not using that, and when successful, the parity is deduced and used afterwards by Ekos in the options, theoretically halving search time.
We're then left with a few options. The FOV of your setup could end up not being what you think: you can use one of your frames directly on nova.astrometry.net and have a solver job provide you with the information. Or your timeout could be too short, although 3 minutes is already very long. My 1.6GHz atom with 1GB of ram usually solves the first frame under a minute.

-Eric
5 years 6 months ago #30566

Please Log in or Create an account to join the conversation.

  • Posts: 407
  • Thank you received: 74

Replied by Clive Stachon on topic Ekos plate solver woes

Just a suggestion but check the number of stars being resolved. If it very high ( i limit it to < 500) you can modify the "down sampling" from the default 2 to say 10 to lower the number of stars used - for me this helps platesolve work 99% of the time and faster.

Note on poor seeing nights you will have to lower the value back down towards 2 depending on the number of stars recognised.

Hope that works for you too !!!
RPI3 Ubuntu 16.04 / AMD desktop Kstars under Ubuntu 16.04 Mounts :azeq6 ,SWAZGoTo

RPI3 Fedora testing out on AMD desktop Fedpra 28 - running kstars 2.9.4 , Indilib 1.7.4 ?????
The following user(s) said Thank You: Jasem Mutlaq, Eric, Jerry Black
5 years 6 months ago #30590

Please Log in or Create an account to join the conversation.

  • Posts: 1029
  • Thank you received: 301

Replied by Eric on topic Ekos plate solver woes

@stash Good point!
5 years 6 months ago #30607

Please Log in or Create an account to join the conversation.

  • Posts: 300
  • Thank you received: 57

Replied by Scott Denning on topic [SOLVED] Ekos plate solver woes

@stash @Eric THAT'S IT!

There were too many stars (I'm imaging in the middle of the Milky Way, in Cygnus).

I downsampled by 10, and my Intel Compute Stick (2 GB RAM) solved in 36 seconds!

I will pay close attention to this in the future. Probably during galaxy season, or when imaging with long focal lengths, I will need to go back to 2, but for now with a wide-angle scope in the MW, downscaling is the way to go!

THANK YOU!
The following user(s) said Thank You: Jasem Mutlaq, Eric, Tarun, Jerry Black
5 years 6 months ago #30610

Please Log in or Create an account to join the conversation.

  • Posts: 300
  • Thank you received: 57

Replied by Scott Denning on topic Re:Ekos plate solver woes


@jerry This is *extremely* helpful! Thank you VERY MUCH!
5 years 5 months ago #30643

Please Log in or Create an account to join the conversation.

Time to create page: 0.388 seconds