Measure magnitude within a circle
Stacking program, astrometric (plate) solver, fits file viewer
Brought to you by:
han59
Menu
▾
▴
At the moment you can only measre the total magnitude within a rectangle. However, a comet coma is generally circular. Can that be added as an option so that we can get a more accurate measure of comet coma magnitude? Also a measure of the diameter of the circle / coma would be useful for reporting.
Thanks
Mark
Last edit: Mark Phillips 2023-03-13
Hi Mark,
Yes that is good idea.
The current setup measures the background just outside the rectangle. This is no problem as long you place the whole object within the rectangle. It will be a little tricky to measure the comet coma only since the tail will be partly used for background measurement. Probably you have to define the place for background measurement separately or better measure in four directions and take the lowest background value. The background of the whole image is less suitable since it could processed without a flat.
For testing, can you provide me with some sample comet images where you measured and documented the coma and comet magnitude? Just for comparison.
Cheers, Han
Just note ASTAP version 2023-2-26 /27 where buggy ASTAP version for the magnitude measurement. Use older or new versions.
Thanks Han
I'm talking to someone who know much more about comet photometry than I do and will get back to you soon.
Regards
Mark
Hi Mark,
I will look into adding an ellipse shaped selection in 4 or 5 days. First I have complete current database change to D05, D20 and D50 for Linux en Mac.
Han
An ellipse shaped selection was added in the development version 2023-03-21. This works for some menus, the magnitude within a selected area, copy past function and two or three other menus.
The background is measured around the original square area. It calculates the median background value so likely a tail will be omitted.
Just hold the CTRL of SHIFT prior to pulling the shape for an ellipse.
It is not possible to have an ellipse under an angle. So it is either horizontal or vertical.
http://www.hnsky.org/astap_setup.exe
Under development.
Han
Thanks will check it out
Yes that seems to work well. Centering the circle around the coma is a little tricky as you have to drag from the corners of an imaginary rectangle. Don't know how that could be improved in some way?
Thanks again from the great work.
Mark
Hi Mark,
About the background used for measuring the comet head. Probably the best thing you could do is to copy & paste the comet head to an other place in the image. It is maybe a little weird but at the same time an elegant way to guarantee the background is measured correctly.
About positioning, I will look into it.
Today the V17 will be replaced by the V50. You will not see a difference but that will be universal for both solving and photometry.
I also saw a request for other colours like B and rc. That is also possible. They can be calculated from Gaia catalog.
Han
Yes that does seem slightly weird but I'll give it a go and see how that works.
There is certainly some interest in using ASTAP in the BAA for comet photometry.
Mark
Not really sure what I'm doing or trying to achieve by copying to a different place in the frame. It's not a very easy process either, to grab just the coma and then grab the copy of it.
Thanks
It is for measuring the head only. If you want to measure the head accurately , you will need a good value for the background. Without moving the head a part of the background is the tail. It is measuring background just outside the original rectangle.
The background is measured using a median algorithm. As long the tail is not covering 50% of the space just outside the rectangle you should be fine. But copying the head away from the tail and measure against a relative clean background it should prove if it is required or not.
See attached sketch. The left head is copied. Inside the orange circle the head is measured. Between the two read rectangles the back ground is measured. For the original comet (right one), 40% of the background is occupied by the tail.
If you just measured the whole comet there is no problem.
Han
Last edit: han.k 2023-03-26
Again the correct sketch. I have to rename it first.
Ah now I see. Thanks for explaining.
I'm going to see if some other BAA people want to get involved in this discussion. People who know more about comet photometry than me.
Thanks
Mark
In order to get a more accurate circle surrounding the coma, I wonder if it's possible to click in the centre of the coma and drag outwards, creating a circle centred on the coma, rather than dragging from the imaginary corners? Don't know it that's easily possible?
That would need some more programming (only circle no ellipses) but which part of the comet exactly do you want to measure? I always understand you need the magnitude of the whole comet. This head measurement must be somewhere defined. There is no clear border between the head and the tail. It is a slow transition with no clear border. Where does the head changes in the the tail?
You can measure the whole comets with a rectangle or an ellipse. The center of the comet head will always be problematic because you need a relative clean background value.
An other problem is stars. Bright stars shining through the head or tail need to be removed for an accurate measurement. That is all possible by copy/paste function. You can stamp the background stars more or less away with the copy & paste function.
For a number of years I have been using Russ Laher's APT (A Photometry Tool) for VS photometry. It is a professional product written by a professional at Palomar for use by both amateurs professionals. Very highly recommended.
APT implements elliptical apertures, designed for galaxy photometry but invaluable for measuring slightly trailed images. It also has superb sky background estimators --- several types to cater for different circumstances --- which are also easily configured.
Not only that, it is fully portable by being wrotten in Java and full source code is available at https://www.aperturephotometry.org/
Russ is also very amenable to receiving bug reports (as would be expected) and requests for enhancement. I have made a couple of the latter which he has gone on to implement.
While I am on the subject and while enhancements are on the table, an aperture which consists of a rectangle with semicircular ends is very useful for asteroidal astrometry where long trails are commonplace. Perhaps this may also be worth considering.
(Added in edit) I will suggest this functionality to Russ.
What I am really saying: learn from others and get in their good books so that they can also learn from you.
Paul
Last edit: Xilman 2023-03-26
Doh! Aperture Photometry Tool
Hi Xilman,
That is an interesting tool to compare. Looking to the video the background seems to be a simple median calculation. At least that is the situation in the old demo video. Automatic flux calibration is not implemented, that is cumbersome.
I thought all professionals where working the fully automatic Sextractor tool?
The first thing I need to understand better how Mark wants to measure the comet head. The tail is in the way for a good background level measurement.
Han
As noted, there are several background estimators in APT. That shown in the video is only one of them.
I really don't know whether APT uses Sextractor (which I use myself, BTW) under the covers. Fairly sure that it doesn't call out to an external utility but it may well have re-implemented its algorithms.
I should find out. Thanks for drawing it to my attention,
Just realised that the flux from the background stars can be automatically subtracted from comets and galaxies down to magnitude 21 using Gaia catalog. That is not so difficult. The star shape is not required. You could just remove flux from star centroid location till all star flux is used.If that is done for the whole image a visual inspection is possible prior to measure the comet magnitude.
I have already a routine which annotates unknown stars (nova) so most code is already there.
han.k: you may already have examined how DAOPHOT does this. If not, the TL;DR is to fit a PSF, which could be either a simple Gaussian or an arbitrarily shaped "blob", to each detected star and then subtract that from the image; repeat until only background remains.
Incidentally, this is a superb though computationally expensive algorithm for background estimation.
That sounds really interesting
Xilman,
It that PSF done for the background estimation?
For most functions I use the median value for background, I have used an iterative mean and sigma clipping for the background estimation but could not find a notable improvement. So this was abandoned. In one or two the functions a "Mode" function (most common) using a histogram is used for background estimation.
The sigma clip measured background is still available in the popup menu "Show statistics in current rectangle/shape".
I found that looking into other programs is at least for me not very productive. It takes a very long time and effort to understand from the code what's happening, if ever. Having the documented principles behind the program function are much more helpful.
I haven't applied a PSF function. Maybe in the future. Maybe it is required for the planned function to remove the background stars. I will know in one or two days.
Since your very familiar with the topic, do you have some challenging objects captured in FITS files with documented measurements to test the photometry of ASTAP? So images containing galaxies where you measured the total magnitude or what else has to be measured? That would be interesting to test.
Thanks, Han
Some (most) of your questions will require more research than an instant answer can realistically provide. I will try to get back to you but I am suffering from a particularly bad cold at the moment (not covid, thankfully) so it may take a few days.
In the meantime, I strongly recommend that you contact Russ Laher --- he is a very friendly and helpful guy --- about how he estimates the background in APT. Please feel free to mention my name (Paul Leyland) if you think it may help, or not if you don't.
I guess the word "DAOPHOT" is familiar to you. If not, it damned well ought to be ;-)
As for test images: I have never really tried to measure galaxies. I can certainly provide images because SV Ari has a superb almost edge-on galaxy very nearby to the VS. AAVSO provides finder charts and accurate comparisons, which I can also provide.