Plate solving issues with crowded field

[Normand Rivard]

Hi Han,

I have many hours of data of M11, a quite crowded field. ASTAP gives me inconsistent result plate solving my images. Out of 143 (mostly good) images, only 80 are solved successfully. I am using latest version (v2025.11.06) with D80 star database.

Here is an example of fits image that doesn't solve.
Any help is appreciated.
--Normand

[han.k]

Hi Normand,

I tried to solve it with several settings but failed. I assume the cause of the problem is the I-filter. In infrared, the magnitudes are very different then in visual. The algorithm will selected as brightest stars which could be much fainter in visual resulting in a mismatch with the database.

The only problem to fix this would be a star database with a passband closer to infrared. That is a possibility I considered for some time for photometry. Currently the standard database contains the Gaia blue magnitudes which is very similar to visual. I have also a local version (V50) containing Johnson-V but that doesn't help for your problem.

Creating a new database takes some effort but I could experiment with the Gaia online feature in ASTAP already available for photometry. This avoids creating an new local database.

I assume you took this image for photometric purposes? For what purpose does these image have to be solved?

cs Han

[Normand Rivard]

Hello Han,
Yes, it is for photometry purpose. I am monitoring variable stars in M11, sometimes in V, and sometimes in Ic.

[han.k]

A quick solution would be use a local Astrometry.net solver and access it via the ASTAP Tools, Batch Processing routine to solve all your images.

[han.k]

Hi Normand,

I have a few images take with a near infrared filter but they all solve well using the standard database. Images of the crowded Milky Way like around M11 could be more difficult to solve especially when they contain many similar in magnitude stars.

But a special star database with infrared magnitude will work better.

As an experiment I have decided to create an special D80 database using the Gaia red magnitudes instead of blue. Gaia red covers 750 - 1000 nm which is very similar Ic passband 750 -900 nm. The processing will take a few days but it goes automatic and I just have to let it run. So in a few days I will know if it helps.

I decided not to use online database version for solving since it is more complicated and it will be very slow.

Could you share a M11 image using a V-filter for comparison? If you have any other I-filter image which fails to solve, can you share them as well? I could use them for testing the new database.

cs, Han

[han.k]

I have created a new database called I80:

https://drive.google.com/file/d/1Rg37PMcgxzYiNh1xkmC1fZWWAGYnzMWN/view?usp=sharing

It solves your image but you have to set the maximum number of stars temporary to around 50 in tab alignment. The default 500 doesn't work because all fainter stars have a very similar magnitude. Select I80 as star database in tab alignment. You also have to install the latest ASTAP from today, I will upload later. Existing versions of ASTAP will not recognise the I80.

I tested some Las Cumbres observatory images in I band and if the image is photometric calibrated in ASTAP, the stars can be measured manually at the mouse cursor in about 0.01 magnitude tolerance.

Maybe not of interest, but I will check if the photometry tab works well with I-filter images.

Han

[han.k]

Today I released only the ASTAP development version for Windows.

http://www.hnsky.org/astap_setup.exe

This version will work with the I80 star database. Are you running under Windows?

cs, Han

[Normand Rivard]

Yes I am. Thanks a lot Han, I will look at that closely.

[Normand Rivard]

Hello Han,
Thanks for the release. I tried it on my Ic data and it worked very well, except for one image. Can't understand why as this image looks OK to me when I eyeball it. Perhaps you can give a try on it and explain why it won't solve.
Cheers.
Normand

[han.k]

Hi Normand,

It required some experimenting/tweaking, but it finds a match if you enter the following:

Radius search radius > 5 degrees
Maximum number of stars to use: 50
Field of view: 0.35 degrees This is essential and forces the solver to user more stars from the database then required. So a higher star density for the star database then the image.

If you annotate the stars you can see the match is correct. Then just save the file so the solution is stored.

The cause is just the abundant amount of equals magnitude stars around M11. I assume it just picks different stars from the database then from the image.

cs, Han