I was Imaging M13 tonight and as usual asked SGP to centre now having initialised with a blind ANSVR solve.
ASTAP would not solve the M13 region - moving to 12 degrees as indicated by the small ASTAP dialog at top left of my screen. Once it became obvious that ASTAP was not solving this one, I swapped to PS2 in the control panel and got a solve within a few seconds and a successful centre of the object.
Perhaps there were too many stars in the cluster for ASTAP?
I thought I’d post this to try to help refine ASTAP, which generally is very fast.
Paul
Yes platesolving could be more challanging if the image is full of stars. Sometimes a shorter exposure could help. What was the exposure time and do you know your field of view?
If you could preserve the image next time and share it, that could be interesting for further analysis.
That is a size where M13 fills the whole image. With that small FOV it will help in any case to install the larger database G18 up to magnitude 18. You can download that one from the ASTAP webpage. Uninstall the G17 since ASTAP will not use the G18 unless the G17 is missing or the G18 is set in the ASTAP menu.
Here is a link to a Jpg version of the file Han. I am using the G18 catalogue, following a previous post where ASTAP was having difficulty in solving my frames. It is better with the G18 Cat., but did not solve the M13 field as mentioned.
I get it solved by activating “slow”. It also helps to increase the quad tolerance from 0.005 to 0.007. The G18 is not required but better for your small field.
Slow means the solver goes with 50% overlap through the sky. For your image it sees 500 stars in the M13 image so it likes to go fast with 0% overlap through the sky database but that is not the best option with M13 creating most star detection’s. It is mabe 30 % slower for blind solving but not for near solves.
To fix it open ASTAP via SGP or directly. Press the Σ button (or ctrl-A), go to tab alignment. Then check mark “slow” and change quad tolerance to 0.007. See attached screenshot. Then to save these setting go to the ASTAP viewer and select the file menu and select exit (and save settings) or just press press ctrl+End to leave the program.( do not close the window because then you will loose the settings)
In SGP settings, solver menu set binning not too high. Binning at 1 or 2 (best).
Thanks for this Han. In your post, you mention blind solving. Can ASTAP be used for blind solving? I thought it needed a sky location hint. When I tried (by mistake) to solve an image in SGP by right click and from the dialog picked solve, i got a message which stated I had to provide further info.
It would be good to know about ASTAP and any blind solving capability.
Thanks for this Han - It would be great to have a ‘check for updates’ feature in ASTAP, or as SGP does a popup to inform users that a new version is available.
best wishes
Paul
If you specify in SGP the max regions ASTAP will search up to 20 or 30 degrees distance before SGP pulls the plug on time. ASTAP, with your field of view and “Slow” does about 62 seconds for a 10 degrees offset and 4 * 62 seconds for 20 degrees offset on my computer.
I still hope the SGP team will use the native ASTAP command line in future where you can specify a search range on degrees rather then regions.
At the moment I don´t have a good solution for automatic popup for update available. But if you download ASTAP from Sourceforge it will ask you to apply for email alerts. If you enter there your email there you get an email from Sourceforge as soon I update the program there.
I had a closer look to the reason why it fails on M13. The reason is that the Gaia database contains more stars of M13 then you can see with an telescope on Earth due to the seeing limitations. It sees the cluster much sharper. This results in the effect that if the program extracts 500 stars from the image and compares them with 500 stars of the Gaia database you get different populations. Matching will only successful for the more distant stars in the neighborhood of M13. The problem only occurs for small field-of-views like with your setup. I don´t have an easy fix for that. The option ‘slow’ helps.
Could give me a copy of the RAW or FITS version of M13 image for further testing? The jpeg is less suitable due to the 8 bits and compression.
Also if you experience a similar problem please send me the image in RAW of FITS format. (you could drop it at ASTAP forum or use Google drive) for further testing. A few users have reported failures in star crowed milky way areas and typically used long focal lengths.
I’m working on a better fix for solving M13 and other global clusters images. The problem is definitely related to the Gaia star database used. It contains several hunderds of M13 center stars at distances of a few arc seconds spoiling pattern match. These stars you can’t image easily from Earth.
The fix seems to work but I’m interested in more M13 (and/or M15, M3) images for testing preferable in RAW of FITs format. The exposure time a few seconds up to maybe 20 seconds. If you could provide me with some samples that would help with further testing.
Ok no probs, I’ll post a link to my 10 sec version. I think I have just the one at present. I have longer exposures if they’d be of use.
I set the ’ slow ’ feature and it seemed to solve the m13 problem.
Best
Paul
This was a major overhaul of the solver. I tried several approaches and found one which works well. You can now switch off the option “slow”
I have considered updating the database and remove the global cluster center stars but the problem is wide spread. Now ASTAP ignores detected database quads with dimensions smaller then found in the image. As a bonus, the solving speed looks a little better due to an improved program flow.
Would be nice if users could test this version on some global clusters and if they have an image where it still fails, please share it. Or just report problem solved
More adjustments where made and ASTAP 0.9.423 is now released on Sourceforge. This version seems to work better with global cluster images extracted from the Deep-Sky-Survey images then PS2.
nova.astrometry.net is unbeatable in reliability. If I crop this image further, ASTAP will fail and nova.astrometry.net not, but it is only 12.8 x 11.4 arcmin and 759 x 676 pixels so I´m happy with the performance.
I have changed the settling time of my mount from 1 to 4 seconds. Curently imaging M27 last two nights and all solves are spot on, centering is fine and happens within one or two iterations at most.
I rough initialise my mount by pointing it CWD and Az 270. Then a slew to M27 got close enough for ASTAP to solve without resort to the Blind Solver, which is new for me. So thanks Han for support with this, it’s working really well. If per chance I get any problems with globular clusters I’ll post back.
Paul
