ASTAP, Astrometric stacking program, looking for ß testers

[han59]

My new free stacking program ASTAP, Astrometric STAcking Program is just released as version ß0.0.7a. I'm looking for beta testers to test the functionality and suggest improvements. It uses FITS files as input and is released currently for MS-Windows only but a Linux version will follow. Conversion from PNG, TIFF to FITS is possible but only in 8/24 bit resolution.

Webpage:

http://www.hnsky.org/astap.htm

 

Description:

ASTAP is a free stacking program for deep sky images. This program can view and stack astronomical FITS images using an astrometric solution. It uses a local version of Astrometry.net or PlateSolve2 to find the astrometric solution. With the solution the astronomical position of each pixel can be calculated. 

Main features:
 

1. Stacking images including Dark Frame and Flat Field correction.
2. Filtering of deep sky images based on HFD value and average value.
3. Automatic call to local Astrometry.net or Platesolve2 for the astrometric solution.
4. FITS viewer with swipe functionality.
5. Results can be saved to 16 bit or float (-32) FITS file
6. Export to 16 or 32 bit integer TIFF files, grayscale or color or simple PNG or JPEG.
7. Mosaic building covering large areas using the astrometric linear solution. 
8. Some pixel math functions and digital development process

 

 

 

 

[Mognet]

I gave it a quick try after seeing your other post earlier, and I just get an access violation when clicking the stack selected images button. Does it output a log file I can send to you so you can see what's causing it?

[han59]

There is no log file. Most likely it has problems with the FITS file format. Viewing should then also be problematic. Just send me one of FITS file you try to stack. I send you my email address in a private message. In the debugger I should be able to see where it goes wrong.

[Mognet]

Sample files sent. Occurs with files converted with PIPP, and another program

[han59]

Thanks for the files. The files can be read without problem. The exception is caused by a missing flat file. It should be able to handle but I forget a check in the code.

I'm trying now to stack them. The images are made with a short focal length around 50 mm. Both local Astrometry.net and PlateSolve2 have a problem with that. Need a few hours to figure this out.

 

 

 

[han59]

For the moment the solvers can't handle a image made with a 35 mm lens. PlateSolve2 doesn't find a solution and the local Astrometry.net finds a solution but gives an error.  So it will work only for longer focal lengths. Anyhow I fixed an error using PlateSolve2 and uploaded version ß0.0.8.

 

[Mognet]

Thanks. I'll download the new version and see if I've got anything with a longer focal length I can test with. Most of what I have is wide field so far

[han59]

You have download one of these:

Installation of astrometry.net (preferred option via ANSVR or  Astrotortilla) or   Platesolve2 including star database as the astrometric solver.  Astrometry.net is a pretty large package with the index files.

For files made with DSLR, having no header with info, you have to enter roughly the 1) position in the viewer and 2) specify the pixel size and focal length in TAB astrometry of the stack menu.  Acquisition programs producing FITS files will have this info already included in the FITS header.

I think it will be good to have a link to an online solver. I'm not sure if this can be done easily. Will have a look to this. In principle you only need the WCS files produced by astrometry.net. It should be possible to request and download them manually but automated will be much more convenient.

Han

Later, the ANSVR link contains a newer compilation of astrometry.net made for ASP. It runs as a Linux program under Cygwin in MSWindows. Follow up to installation step 9. The download of the index files was pretty quick done in less 15 minutes. The link you have to put in ASTAP is as follows:

C:\Users\user_name\AppData\Local\cygwin_ansvr\bin\bash.exe

Adapt "user_name" to the login name used in Windows.

It will still not solve an image of 37.2 x 24.6 degrees. The index files run up to 33 degrees. It solves but then stop with an error.

 

 

 

[Mognet]

Thanks Han

I've installed AstroTortilla and PlateSolve2, but it may be that the subs I have taken at longer focal lengths are not of good enough quality. There are clear skies forecast here for tonight, so I will try to get some better ones. I will try solving with ANSVR too.

Astrometry.net has an API that might be worth investigating. There are some details at http://nova.astrometry.net/api_help

 

[han59]

I had a brief look to the API. To get that working will need some time.

The images don't need to be of good quality. For astrometry.net switch in the in TAB astrometry the pixel size limitation off and set the position range to 10 degrees or ever higher. That should still give a quick solving. It could that be that the OSC images are more problematic. Anyhow if you can't solve, send me one or two pictures and I will have a look. It is a good test case.

As an experiment you could just upload an image to nova.astrometry.net without any information. The online solver is very good and will solve almost everything. This will give some clue of the position and pixelsize.

Anyhow astrometric/plate solving doesn't seems to be easy for everybody. But as soon it is working, it is of great use for pointing your telescope (unless your just use wide field and can see everything)

 

[Mognet]

ANSVR works for me. Now I just need some decent subs for stacking. The 35mm lens is too wide for plate solving, so I'll try with the 18-55 kit lens at 55mm

[han59]

Did some testing this morning with OSC images from my QHY8 and noticed Bayer conversion for color is not applied if astrometry.net is used. Is fixed in version ß0.0.8a

 

[Mognet]

I've downloaded the latest version, and the images are now appearing in colour even when looking at a single unprocessed image. I had assumed it was an issue with the source images as some other stacking programs had also displayed them in monochrome.

Still waiting for that elusive clear sky so I can get some suitable images to test properly with

[han59]

I managed to solve the two images you send me and then stacking was then easy. With an astrometry.net  setting for downsampling and it can solve the images.  It requires a little tweaking and each image is seen as different by astrometry.net. I did it with the following two commands:

 --downsample 2  --sigma 10

and for the orther image as:

 --downsample 9

Alternatively you could solve them online and stack the result using the produced .new files.

In the new ASTAP version ß0.1.1 you can keep the console open and see what for error the solver astrometry.net error is reporting. The stack of the two images provided is not good. The field is too distorted and alignment is poor. A focal length of 30 mm is no success. Longer focal lengths are required. Most likely the zoom lens at 55 mm will also give problems.  Anyhow the program made significant progress. Also the SIP polynomials from the solution are included. The should help but not in all cases.

 

2 images of 30mm focal length stacked:

 

 

[Guest]

Hello from Vienna, I also did download the latest version and need or kindly ask for a clarification, where to install the (ASTAP.EXE) file ? under the link http://www.hnsky.org/astap.htm it is statated to..

1) Install a local copy of Astrometry.net (preferred option via ANSVR  or  Astrotortilla) or   Platesolve2 including star database as the astrometric solver. ( which I already did ) but where should I install or store the ASTAP.EXE file ? can I store that any where I wish ? or should I staor the ASTAP.EXE in the directory where one of the apps ANSVR  or  Astrotortilla or   Platesolve2 is installed ??

 

Thank you and all the best

 

Martin

[han59]

Hello Martin,

You could place the program astap.exe in any directory or map. It could be under c:\astap  or  c:\Program Files\astap or simple on your desktop. Under c:\Program Files\astap would be the most appropriate.

For ANSVR the CYGWIN directory will be at    C:\Users\user_name\AppData\Local\cygwin_ansvr\bin\bash.exe

For Astrotortilla as c:\Cygwin\bin\bash.exe

Both will install a 2010 version of Astrometry.net and will perform equally.

There are two new options in the stackmenu. Set it as follows:

• Show solving show the Astrometry.net solving console. You better keep this on to see what happening.
• Keep console open: This should be off and is only required to see an solving error.
• Search area: This can be large 10 degrees or so.
• Pixel tolerance: Can be off.
• Extra options: Can be blank. --Downsample 2 or --Downsample 4 helps sometimes with solving. For noisy images you could experiment with --sigma 10  or higher.

For very large images Astrometry.net seems to provide a more accurate orientation angle and therefore result.

I like to hear how it performs. You could also join the Yahoo group just started up or send me directly a mail.

Han

[sloz1664]

Can ASTAP be used with Pinpoint plate solver?

Steve

[han59]

No you can't.  Pinpoint is a commercial product. I didn't investigate using it. I know that Pointpoint Light can't be accessed remotely so only the full version could be used, but since there free solvers available there is no need to use it.

Han

[sloz1664]

Thanks Han.

Steve

[Guest]

Hello Han, i did already request to join Yahoo group this morning, no anwer yet..

I did download the latest version of ASTAP, and had few carashes at the start but finally, stacking incl. Plate solving with platesolve2 did work.. with Ansvr not! did fail to solve images, Platesolve2 works  much better

I will for sure use ASTAP from now on and follow the progress of developement.. 

Regards 

Martin

[Guest]

Hi Han, Yahoo Group notification well recieved few minutes ago.. thank you, Martin

[han59]

Martin,

I didn't see a message from Yahoo that your where pending in the group. I just approved you. Maybe something to configure.

PlateSolve2 worked pretty good but for one test series, I noticed that the stars in the corners became very small ellipses due to a small angle variation in the solutions. I could not fix that. Keep an eye on that . If the same problem occurs with your images you need to get Astrometry.net working. For the rest PlateSolve2 works fine and fast.

 

 

 

[han59]

ASTAP is now at version ß0.2.1. Below a mosaic x 4,  I created today in ASTAP using 154 images of 50 seconds, so average about 1/2 hour for each tile. They are not equal distributed. The south part is noisier since there where not enough images available:

Object: M31
Date: 2017-11-18 & 22
Telescope 100 mm APO Imaging Star astrograph APO100Q, F5,8
camera ASI1600MM-Cool, bin x2

4 tiles of total 154 images using CCDciel for imaging.

Work method:

1. stacked sigma mean using local Astrometry.net for solving (option --downsampling 2)
2. Mosaic option ON, oversize 2500 pixels.
3. Cropped stacked result to about 5000x5000 pixels
4. Applied median-equalising filter under Pixel Math
5. Applied DDP (digital development process) under Pixel Math
6. Adjusted range and saved as JPEG, 90% quality

Link to full size mosaic

 

[han59]

The first Linux version of ASTAP is ready. If something is not working as expected, please inform me.

Some installation instructions are provide at the web page. You have to install a local version of Astrometry.net (preferred) or use Wine & PlateSolve2 for the solving. In principle you could also use the online version of Astrometry.net manually to get the solution in a .wcs file or as a new file with an updated header but that is very cumbersome. The solution in the FITS file provided by the acquisition program is in almost all cases the target value and not accurate and should be overruled in the program.

The MS-Windows version is also updated and both versions are now at version ß0.4.0

Han