Warning, remaining image dimensions too low!
Stacking program, astrometric (plate) solver, fits file viewer
Brought to you by:
han59
Menu
▾
▴
Hi all,
I am running a series of 85 mm images from a dense southern milky way region taken with a Nikon D5100 (APS-C format, attached). FOV is set to 10°, radius search area to 180° and downsampling set to 0.
I am getting the following warning:
23:19:18 Using star database V50
23:19:18 Large FOV, use G05 database!
23:19:18 Cropping image x 0.49
23:19:18 █ █ █ █ █ █ Warning, remaining image dimensions too low! Try to REDUCE OR REMOVE DOWNSAMPLING. Set this option in stack menu, tab alignment.
23:19:18 525 stars, 425 quads selected in the image. 348 database stars, 282 database quads required for the 5.14° square search window. Step size 5.14°.
23:19:22 No solution found! :(
Obviously I cannot reduce downsampling any further. At 50% reduction the pixelsize would still be about 2500 x 1600 which I understand should be enough to solve the image. Using the G05 database solves only some individual images but fails for most images. For my purpose (photometry) the V50 database would be preferable.
Any ideas how to handle these images?
cs Axel
Hi Axel,
Yes that image is difficult to solve. Huge amount of stars having similar magnitudes.
Best solution now will be to use the older V17 database. That one has tiles of 10 degrees instead of 5 degrees and that will avoid the image cropping. It is also intended for photometry. So it contains the Johnson-V magnitude like the V05 but limited by mangitude and not star density.
The V17 you can download from here:
https://sourceforge.net/projects/hnsky/files/star_databases/v17_star_database_mag17_colour.zip/download
It is a zipped file. Just unzip into the ASTAP directory/folder. After installing select in the ASTAP settings, tab alignment the V17 as star database. The V50 (5000 stars/degrees) you could remove manually or uninstall.
Next weeks I will make a new V05 stardatabase. Same like the G05 with a star density of 500 stars/degrees but as the star magnitude calculated Johnson-V instead of Gaia BM mangitude. The older V17 is limited by magnitude 17.
Han
Last edit: han.k 2024-07-12
Correction the V05 is already available. I will change the message to
You can just download an installer:
Windows:
https://sourceforge.net/projects/astap-program/files/star_databases/v05_star_database.exe/download
Linux:
https://sourceforge.net/projects/astap-program/files/star_databases/v05_star_database.deb/download
Mac:
https://sourceforge.net/projects/astap-program/files/star_databases/v05_star_database.pkg/download
This V05 solves your image.
Han
Han,
thanks, that helped to some extent. I downloaded and installed the V05 database (and a new variable star database, as proposed by the program). This solves about 80% of the images, some 30 out of 150 images remain unsolved. I have attached an unsolved image.
Trying a photometric calibration on the first solved image gives a calibration failure:
Photometric calibration of the measured stellar flux.
Median HFD is 4.3. Aperture and annulus will be adapted accordingly.
Using star database V05
Calibration failure! Less then three usable stars found.
Do you have an idea how to solve these problems?
Axel
Han,
I tried another similar series (85 mm, 10x15° fov, southern milky way field) that was solved without problems with V05 but would not accept a photometric calibration.
3:24:45 G:\astro\Veränderliche\Mus\Mus TU\20230611\lc\lc.cmpack-files\tmp00001.fts Adding astrometric solution to file.
13:24:45 Using star database V05
13:24:45 Cropping image x 0.90
13:24:46 1023 stars, 794 quads selected in the image. 678 database stars, 526 database quads required for the 9.53° square search window. Step size 9.53°.
13:24:46 34 of 35 quads selected matching within 0.007 tolerance. Solution["] x:=22.921993x+ 3.317976y+ -31073.573341, y:=3.323348x+ -22.925211y+ 14676.209243
13:24:46 Solution found: 11: 21 02.83 -64° 25 30.9 Solved in 0.6 sec. Δ was 13.3". Used stars down to magnitude: 9.4
13:24:56 Can not calibrate G:\astro\Veränderliche\Mus\Mus TU\20230611\lc\lc.cmpack-files\tmp00001.fts. Add first an astrometrical solution.
When viewing the image another problem occurs: the image is completely covered with annotations (see attachment). Zooming in does not help as the font size remains unchanged. Trying to find a specific variable (with marker at a,d position - is there a better option to zoom to a star?) is quite impossible . How to get photometric measurements in this case?
cs Axel
Hi Axel,
That is a very crowded area. I have made an artificial image of the same area and it is not workable. Probably I have to limit the variable star density. Is there a specific variable star your interested in?
Can you share one of the images?
Han
That is weird. Share one test image.
This is the field down to magnitude 9:
Hi Han,
I am usually centering the images on one short-period var. (here TU Mus) but I am also searching the field for other short period variables that might have a minimum/maximum during the observation period. I then stack the images and get the magnitudes for long period variables from the stacked image.
ASTAP is a great photometry tool and getting these crowded fields analyzed would be extremely helpful!
cs Axel
Hi Alex,
The reason it won't do photometric calibration is that most stars are saturated. ASTAP detects that the maximum value in the image is 17721 and only allow stars with a peak value of 16721 := 17721-1000. So for your setup it would require a shorter exposure time then 30 seconds.
The image is pretty undersampled. Is this attached image a binned version?
This field seems to be highly investigated and reported. I see two ways to reduce the number of annotations. Reduce the limiting magnitude to maybe magnitude 7 or 8. Or create a database of short period variables only or a filter on online version. That would remove maybe half of the variables.
Looking to the output of AAVSO, I see also the epoch and period of the variable. Could that be used to filter?:
E.g. SS LEo:
https://www.aavso.org/vsx/index.php?view=api.list&ra=173.478667&dec=-0.033698&radius=0.350582&tomag=13.0000&format=json
The epoch is in the year 1973 so pretty old:
VariabilityType "RRAB"
Period "0.6263351"
Epoch "2441781.409"
RiseDuration "12"
MaxMag "10.420 V"
MinMag "11.584 V"
The question is can epoch and period be used to calculated to min max moment?
How do you know a variable is around max or min?
Han
In the AAVSO VSX database you can generate ephemeris. So for SS Leo, I get this:
Ephemeris for SS Leo (HJD/UT)
Epoch Mid
2460509.455 17 Jul 2024 22:54
2460510.081 18 Jul 2024 13:56
2460510.707 19 Jul 2024 04:58
2460511.334 19 Jul 2024 20:00
2460511.960 20 Jul 2024 11:02
2460512.587 21 Jul 2024 02:04
2460513.213 21 Jul 2024 17:06
2460513.839 22 Jul 2024 08:08
2460514.466 22 Jul 2024 23:10
2460515.092 23 Jul 2024 14:12
2460515.718 24 Jul 2024 05:14
2460516.345 24 Jul 2024 20:16
2460516.971 25 Jul 2024 11:18
2460517.597 26 Jul 2024 02:19
The question is this min or max and is it accurate?
Epoch: 2441781.409
Period: 0.6263351
Epoch by AAVSO calculated: 2460509.455
Nr of periods calculated 29901.00028 := (2460509.455-2441781.409)/0.6263351
Okay looks I can reproduce that. So this could be used to filter out stars near the calculated epoch. Would that be useful?
The image is an original, no binning. All of my southern skies images have been taken with the same settings, so the first image in this thread should show the same problems (saturation) but obviously did not. How can this be explained?
Setting the limiting magnitude for annotations would be very helpful. Is there such an option?
if the ephemerides are given for Min or Max depends on the variable type: max is determined for pulsation type vars, min for eclipsing type vars. There are often a number of elements from different sources and the VSX often contains the most up to date, but not always.
Creating a database of short period variables only or a filter on the online version sounds like a good idea. Filtering near a calculated min/max would also be an interesting option but the most interesting variables show changes in period so their min/max may not be near the calculated time.
Is there a description of the variable database format? Or how to create that filter?
cs Axel
By the way, there is an easy method to detect variable stars in a field which is basically a plot of mean magnitude vs. standard deviation of magnitudes of each star in an image series. If this sounds interesting I could give a more detailed explanation / specification in the feature suggestion forum.
Yes that also crossed my mind. In principle that is already done and reported for the four most variable stars. To help with annotation the program should go through the list prior to annotation. Or this is done as a filter for the "measure all" option. Now you will get a few hundred columns after "measure all"
It looks for at least three none saturated stars. Maybe the first images where a little out of focus or the transparency was a little less just enough to get a pass from the software.
Can you share a few more images of this serie?
The limiting magnitude setting for star annotation is In the photometry tab but the current options ares not sufficient for this crowded field. And by limiting the magnitude you will miss the annotations of the fainter stars. I'm looking for a better option. You could try option "measure all" then you don't have to select the stars manually. With "measure all" you can select any star in the AAVSO report.
The database format is pretty simple. It is .csv text (can be loaded in a spreadsheet) and the format is described in the second line. But at the moment, I think it is better to build a new filter in the program.
Han
See attached. TU Mus is selectable in the ASTAP AAVSO report. The long drop down list is sorted alphabetically. If it would be sorted on standard deviation then you don't need the annotations....
Can you share the full series to test this concept?
By selecting Tu Mus in the report (measure all option) then TU Mus is marked with a purple diamond. See attached.
That is already an excellent option. How are comparison /check stars selected?
Unknown variables are however not detected this way ( I had two already in other image series).
One option would be to restrict the annotations to an area around a given star or coordinate
The complete series are about 90 images (in the last images the horizon begins to intrude from the side). Will send them a s soon as possible.
It can suggest a check star with a measured magnitude closest to the measured variable magnitude.
It will always report the four stars with the highest standard deviation in the log and annotated them with yellow circles. There is no cross check with AAVSO database at the moment. I'm not aware anyone used this feature up to now.
One way to cope with it is to measure the 500 brightest stars. Never tried. There could be a limitation to the number columns but 500 stars looks feasible. But you could also look to Gaia database. It will likely have detected variation already.
It late here. Tommorow there is an other day.
You should have received a WeTransfer email with the link to the files.
Hi Axel,
I did not receive the WeTransfer email. Send the link to the email address I send by personal email.
Cheers, Han
Han, just tried to process my NEF images: loaded NEF images (camera dark corrected) into light and flats tabs, chose stack method calibration only, processing as OSC images, auto Bayer pattern, LibRaw full active area but received an error (see attachment). I guess I am too dumb to do a simple calibration :-( Any idea what to do different?
cs Axel
Strange error. I have changed the procedure involved from inline to normal and it seems gone.
I will now work on a version to sort the all the measured variable on standard deviation. Then it is time to release an version for all operating systems.
No idea. Make a small test set for me to reproduce the error. In the debugger I will quickly find the problem. Send the link to my personal email.
I have just finished the problem with your +32 files. Took some time but now I can work on the annotation again.
Han