Jump to content

548140465_Animationchallenge.jpg.32379dfa6f3bf4bba537689690df680e.jpg

Plate solving and goto with AstroTortilla


RogerTheDodger
 Share

Recommended Posts

You should have a link to Cygwin terminal somewhere. Run that, type solve-field and press enter. That will give you a list of all the switches you can use with a little description.

Taking Themos as an example:

-O : Overwrite existing files

-z 2 : Downsample the image by a factor 2 (can be any integer factor you want)

-u app : This selects the image scale units. app is for arcseconds per pixel

-L 120 : Lower bound of image scale estimate

-H 130 : Upper bound of image scale estimate

Edited by russellhq
Link to comment
Share on other sites

Thanks - yes I have seen that link before. But it does not tell you what happens - for instance, what does downscale or Sigma do and some examples of its use? I have fiddled with these values and got it to work by adjusting them. Maybe I am being a little unrealistic.. The software is superb and I am very grateful to the developers.

I shall have a look into that cygwin environment. I need to dust down my Unix books a little I think - I used to be a Sun Solaris admin in a past life - one forgets these things!

Link to comment
Share on other sites

There's two phases to the software: one is source extraction and the other is matching. The downscale and sigma options affect the first phase, I think. It would be nice for someone to come up with examples of how each one helps in the presence of image artefacts.

Link to comment
Share on other sites

I shall have a look into that cygwin environment. I need to dust down my Unix books a little I think - I used to be a Sun Solaris admin in a past life - one forgets these things!

I still am! Doh brain in weekend mode... :huh:

No man pages but try this..

$ solve-field -help

Link to comment
Share on other sites

I find that arcseconds per pixel has the highest solve rate for me. It should also be quite easy to work out, ( Sensor Pixel Size (um) / Focal Length (mm) ) * 206.3

I then set the lower and upper bound 1 unit lower/higher. I don't normally need to then increase this as it usually solves straight away.

I thought I had all the index files downloaded, turns out I hadn't. That's probably why it wouldn't solve for me!

Although I still don't understand why the solver said the image scale was 127 arcsec/pixel as the calc above says it should be 55 :confused:

Link to comment
Share on other sites

It is worth checking the updated FAQ on the AT sourceforge page. They have added some guidance on additional options to try since this month's V0.5 update. (The parameters have always been there in solve-field, but now the AT guys have pointed out some useful ones I missed before).

I don't have the software on this machine, and our web-filter blocks sourceforge, so I won't guess at the settings, but in summary:

- There is a parameter to re-sort the sources (stars) found by brightness, thus solving for the brightest 10 first, etc. This seems to help a lot, since it ignores most of the noise rather than treating it as small stars.

- There is a parameter to limit the number of sources used to solve. Most of us have been using the sigma parameter to cut down on the number of stars, but actually that may not be optimal and perhaps using a slightly lower sigma and a hard limit on the number of stars would be more reliable, especially if using the re-sort option as well since you be pretty sure you have the best stars in the image to start the solve.

- Also a parameter to relax the limits on the shape of stars. It was not at all obvious what it does from the name of the parameter, but if you aren't getting many sources detected this will help (e.g. poor focus, a bit of coma, etc.)

They also recommend setting the search radius to 45 degrees (only affects live solves, not blind solves from loaded images I think) since apparently most scope drivers won't do a sync if it is off the current position by more than 45 degrees, so it will deduce that it hasn't got a clue where you are pointing in far less time if that is the case (and if you are that far out of alignment to start, you have bigger problems to solve first)!

I tried the new settings on one of my saved images, and cut the solve time down from an acceptable 18s to about 6.5s, so definitely worth tinkering a bit.

Link to comment
Share on other sites

  • 2 weeks later...

Great thread - thank you folks :)

I've just installed AT and ready to "play". Installed sucessfully on the netbook I use for astro imaging, with index files 4005-4019. I've also installed it on my desktop (Win7 64bit OS) which is more convenient for testing indoors with all this bad weather we're getting. Couldn't find the location of the index files to copy across so I'm downloading them again. Working out which index files are wanted for a particular scope/camera combination seems a bit confusing but I think ED80 at 600mm and Atik 314L+ wants 4005. I have a goodly number of image files I can try for plate solving and I'm hoping to adjust parameters and settings to optimise solving.

Link to comment
Share on other sites

Welcome to the world of AT Gina. I think you'll find you're in for a treat! :-) It makes life so much easier!

Sent from my GT-I9100P using Tapatalk 2

Thank you :) I have it working solving images I took a while back and now I'm refining my settings. Now if only we could have some clear night skies...
Link to comment
Share on other sites

Beginning to get my setup right now :) Took the quick test exposure of the heart of The Heart Nebula from my last imaging run, uploaded and solved in astrometry.net which gave 2.56 arcsec/pixel. Then set AT to 2-3 app and AT solved it in 30s on my netbook. (netbook set up in warm room controlling it from living room indoors with TeamViewer). So that's the settings for 314l+ and ED980 plus 0.85x FR :)

Link to comment
Share on other sites

AT solved it in 30s on my netbook.

Hi Gina,

I'm curious as to how you got your netbook to solve so quickly. I'm using an EeePC 1000 and the fastest I've been able to solve is a little over 70s. My setup has an imagescale of 1.79"/pixel and I set the lower limit to 1.78 and 1.8 for the upper limit. I run the solve on my desktop PC first so that I know which fits file the solution is found in and use only that fits file on the netbook so it doesn't have to search through lots of files before solving.

The vast majority of the time is used up by AT doing it's 'writing ppm, using 8-bit output, extracting sources, etc' - this takes well over a minute. The actual solve only takes a few seconds as it has only the correct fits file to check.

So, this leads me to think that your netbook is processing the preliminary stages much more quickly. My netbook is a 1.6GHz Atom with 2GB RAM. Is yours substantially faster? Did you have to tweak some settings to get through the preliminary stages more quickly?

Link to comment
Share on other sites

I use AT on both my ED80 and RC8 scopes and it works brilliantly! That said we have had such little time under the stars this winter - I'd like to do more testing but I begrudge wasting time under the stars for "testing" such is the premium on clear skies... I'd rather be observing or imaging!

Link to comment
Share on other sites

Thanks Russell, I wasn't doing any downscaling but have just tested with downscaling of 2, 4 and 8. It seems 4 is the optimum value for me and has cut my solve time down to 48s. It's a real improvement on what it was but it's still nowhere near Gina's 18s. Hopefully Gina will shed some light on how she managed to achieve that speed.

My custom options are --sigma 10 --no-plots -N none -r -c 0.02

Link to comment
Share on other sites

I change the sigma parameter depending on how many stars are in the image that will be plate solved. However, I find plate solving works fine with as few as 15 stars so I rarely need to change it from 4. On one occasion with my RC8 - reason escapes me - I had to put the sigma to 10 since at 4 there were too few stars. 10 found more stars to solve on I recall.

Link to comment
Share on other sites

Thanks Russell, I wasn't doing any downscaling but have just tested with downscaling of 2, 4 and 8. It seems 4 is the optimum value for me and has cut my solve time down to 48s. It's a real improvement on what it was but it's still nowhere near Gina's 18s. Hopefully Gina will shed some light on how she managed to achieve that speed.

My custom options are --sigma 10 --no-plots -N none -r -c 0.02

How many sources does AT find when you ran that test? Try changing the sigma so it finds around 100. That'll speed things up too if you're finding a lot of sources.

Link to comment
Share on other sites

At sigma 10 it was finding 600ish sources, now with sigma 25 it's finding 130ish sources. Still solving in ~48s.

I've just realised that the most likely reason for Gina's solving being so much faster is the fact that she has far fewer pixels in her Atik 314L+ than I do on a 450D. Approximately 8 times less. ~12Mpixels on 450D, ~1.5Mpixels on Atik 314L+.

<50s is reasonable enough I think. I don't intend to use AT on more than a couple of targets in a session. I'm more than happy to instruct AT to take me to a target, go make a cuppa and come back in 5 minutes to a perfectly centred target. :)

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.