Is anyone, anyone at all, using any software tools that will perform point-spread-function (psf) fitting photometry?
A good number of stars I have observed over the last few years have faint companions, especially in the galactic plane region. The photometry could be improved by removing the blended companion.
Back in the day, decades ago DAOphot was the thing. Today there seems to be nothing available. Aperture Photometry Tool might be able to do it… building the psf kernal, sometimes called the point-response-function (prf) is the hardest part.
I believe siril does a psf fit, as well as aperture photometry. I tried comparing both approaches on some gmcep measures I did, but did not see appreciable differences. I stuck with aperture since the psf required setting an area of interest rather than point and click ( although the aoi seemed arbitrary it worked well and the documentation claimed it better). You can check siril documentation to verify.
DAOphot is still around but unix/linux - yes even solaris! I don’t run any computers that use linux anymore and too old to start again.
Source code is available for DAOPhot for those with a good fortran compiler. I have a decent 32-bit fortran compiler but I know there will be strange dependencies that will cause great frustration to myself so that is out.
Pixinsight might be able to do it but the tools seem to be optimized for generating deconvolution kernals to make more pretty pictures.
I own Pixinsight but the interface is very poor IMHO and after a few steps in is incomprehesible. Siril also has a new-fangled interface that one may get lost in quickly.
Python is a snake isn’t it?
I know Aperture Photometry Tool will do the psf photometry but then I’m left with how do I generate the psf kernal. So I’m down to is there a tool that can pass over an image, generate an average psf kernal of say 25 pixels in size and can then allow the psf kernal to be saved as a fits? APT can then be used to import the kernal and do the fitting and photometry.
Anyone in the AAVSO able to build a psf kernal piece of software that has a basic user interface for Windows 11 Pro? We can probably find in the literature a proper method to generate the average psf kernal.
I’d be your best friend most likely if someone would write the software to generate the psf kernal.
I’ll take a look at siril again but time is a wasting at my age…
I get the “age-time wasting” thing completely! LOL, but have given up on my " the runway is getting too short for this" excuse, given it seems one has to take five sidetracks to get to where you want to be nowadays. We can do a google meet or other, and I can show the psf/photometry in siril on your data, if your other avenues fail and you desire.
Thanks but I am currently trying, evaluating another piece-of-software. Not sure at this point if it is worth the effort. Aperture photometry does 99% of what I need but there is always at least blended variable that might benefit from that extra effort. Thanks to all, and to all a good night… gonna be cold and windy here! Yoo Hoo!
The PSF Photometry program I have been working on has been updated so that one can install it easily using a setup.exe. From the release notes:
Welcome to MAOPhot 1.1.0, a PSF Photometry tool using Astropy 6.1.6 and Photutils 2.0.2 (Presently this has only been tested on Windows 10 and Windows 11).
Important Update: Install MAOPhot by downloading MAOPhot_SETUP.exe and executing. There is no longer a need to download Visual C++ libraries. (This has only been tested on Windows 10 and Windows 11.)
New Feature: “Selected Stars” The list of peaks for generating the effective PSF Model can be examined more closely. The list can be edited.
Welcome to MAOPhot 1.1.4, a PSF Photometry tool using Astropy 6.1.6 and Photutils 2.0.2
MAOPhot 1.1.4 Changes
Support for APASS DR10 [along with AAVSO, GAIA DR2, and APASS DR9]
Added Setting: Max qfit , [‘qfit’ is quality of PSF fit, lower number is better fit; MAOPhot discards any PSF fit with qfit > Max qfit”; anything below 0.1 is considered good]
Added Setting Min Separation Factor [this x FWHM = the minimum distance (in pixels) such that any two sources separated by less than this distance will be placed in the same group]
Added Setting: “From Fits” checkbox for CCD Filter; when unchecked user can manually override FILTER value in Fits header
“Find Peaks” and “Iterative PSF Photometry” now uses DAOStarFinder
In APASS DR10, remove any entries with Johnson (V) > maglimit until cgi-bin/apass_dr10_download.pl is fixed
Use mouse wheel and shift-mouse wheel to scroll
Critical changes
insure that all fitted PSF’s have qfit (Quality of Merit) < “Max qfit” ; Only the lowest qfit match found within the “Matching Radius” is used. This match may not be found until a later iteration of “Iterative PSF Photometry”.
use of APASS DR10. The “Maximum Ensemble Magnitude”, in Settings, is used to limit the inital comps received from the AAVSO. for a FOV > 60’, a mag limit of 15 is recommended, FOV > 120’, a mag limit of 12 is recommended
Due to a lack of available “cradle to grave” photometric pipeline software which uses Iteratively Subtracting PSF Photometry (ISPSF), almost essential for crowded star fields, I ended up writing my own system, ASTROPHOT. It’s working well although I run it on a native Linux server due to the heavy computing cost associated with the process. It’s written in python but I haven’t tried it running under Windows say using an Ubuntu hosted OS running in VirtualBox. I would have said I don’t see any reason why not but there tends to be significant unforseen issues porting anything of this complexity. The ISPSF process is highly parameterised which can be both a blessing and a curse. It is excellent for handling crowded star fields and adapts easily to different use-cases. I am using it to bulk extract stellar objects exhibiting several classes of variability with a periodogram based multi-period detection algorithm dealing witg the data post PSF extraction. I’d be happy to give you the benefit of my experience if that would help.
Gordon
MAOPhot calculates stellar magnitudes from FITS formatted digital photographs using PSF photometry. It produces an extended AAVSO (American Association of Variable Star Observers) report (AAVSO Extended File Format | aavso) which can be submitted to AAVSO using their online tool WebObs (WebObs | aavso).
MAOPhot uses PSF (point spread function) photometry exclusively.
Now that I have no observing site anymore but gazillions of archived images I can try to use MAOPhot on them to compare to the synthetic aperture photometry I original created. I also have at least one very tough blended case worth testing your software on.
If you have a GitHub account you are welcome to enter any new issues concerning MAOPhot. GitHub · Where software is built I will try to resolve them ASAP
Running without reading the pdf always brings things out!
First I have never liked astrometry.net. At least for real single filter CCD-based images ASTAP works really well on my images. The plate solution is written to fits header. Will MAOPhot use the plate solution in the fits header and I can skip the astrometry.net plate solve step?
In your pdf a section with a few examples of good and bad star psfs as examples of stars to “reject” or keep from the ePSF kernal generation would be a good add for some perhaps.
We may have to agree to disagree on he merits of astrometry.net. I find it invaluable and feed almost every image I take through it. I also use it to determine the direction in which the scope is pointing when starting a session. The procedure is to take a snap shot of a random bit of sky, plate solve, synchronize the scope in MaximDL.
I have, of course, installed it locally on a number of machines, most Linux systems but also on the Windows-10 TCS in the guise of ANSRVR. All use the GAIA caatalogues. No hanging around for slow network transfers or sharing an overloaded server with other users.
@petefleurant — MAOPhot looks very impressive and I would like to give it a test run. However, the docs says that it is for Windows 10 and 11 whereas I am in an almost exclusive Linux environment.
Also, it says that the full source code is available, which suggests that it should be able to run just fine on any system with a Python installation.
In principle it may be possible to run under WINE but my experiences with that approach have been very mixed, with usability ranging from seamless to impossible and a number of cases in between.
Could you elaborate, please, before I dive down the rabbit hole?
Sorry for the confusion. What is only supported by windows is the MAOPhot_SETUP.exe.
If you get the source code, you should be able to run MAOPhot.py in python, just make sure you install the required packages as listed in requirements.txt ($pip install -r ./requirements.txt)
