solve fails due to field shift?
Stacking program, astrometric (plate) solver, fits file viewer
Brought to you by:
han59
Menu
▾
▴
Hi all,
I am working on a 175 image series taken for photometric purposes with full-field DSLR and 135 mm lens. Due to bad mount alignment (south pole could not be not properly identified) the field shifted about 5° during the 9 hour session.
When trying to solve the images (astrometric alignment) only the first 50 images were solved. For these solutions astap automatically selected the G05 instead the V50 data base as specified because of the large FOV and cropped the image by 0.5x.
For each subsequent image astap reduced the cropping factor successively down to a 0.4° FOV but could find no solution.
I assumed that astap solves all images independently but this looks as if the field shift has an influence on the solving process. Has anybody experienced similar issues? Any hints how to proceed?
cs Axel
The shift should not influence the solving. It must be something with the stars. Just attach an image which fails to solve and I will have a look.
For stacking the standard "Star alignment " works faster then the "Astrometric alignment". I would default stack with star alignment. Use astrometric only for rare cases where star alignment doesn't work.
Han
Maybe your search radius is less then 5 degrees?
I tried both astrometric and star alignment with 10° search radius, both with similar results. I don't see any differences in the images that solve and that fail to solve. However, in both cases many star images are quite small.
I have uploaded two images to
http://geoatwork.de/downloads/astro/astap-files.zip
tmp00001.fts worked, tmp00050.fts did not. Please have a look.
cs Axel
Hi Axel,
For the image 50 the solver can only match a part of the stars. See yellow marked stars in attachment. Very strange and this required a further investigation. The stars are available since I can annotate them with orange circle but not in the corners.
I will look into this problem later. First I have to finish something else.
Can you try star alignment for the moment? And set the Hash quad tolerance to 0.007? That helps in solving/matching.
Can you a few more images of the series with a different drift? Maybe I can learn something from it.
Cheers, Han
I have the impression there is something wrong with the database tile selection. It uses only one tile.....
Han,
I tried star alignment again, hash code tolerance was by default already set to 0.007. The process was terminated with a 'division by zero' warning which is weird as I had tried that option earlier without any error messages.
I have uploaded some image distributed over the whole series (same URL as before). In the last images the horizon begins to intrude - not sure how this is handled.
I have also added two raw images (corresponding to tmp00001.fts and tmp00050.fts. I had some difficulties calibrating raw images in astap (see the last entry of my post https://sourceforge.net/p/astap-program/discussion/general/thread/ce60cae1d0/ ) and calibrated the images in another program but I guess that should not be an issue.
Glad if you can find the time to have a look.
cs Axel
Fixed in the development version:
http://www.hnsky.org/astap_setup.exe
Tell me if this version works for you . I will update the Linux or Mac version later in the day.
will check as soon as possible
Han,
thanks, this worked far better than the previous version. Only in the last part of the series some images failed (see attachment). Beginning from image 160 the horizon intrudes from the right but among those images there are as much images that have a solution as those that fail. If you like I could provide a set of images that have both worked and failed so you might be able to track down the problem.
cs Axel
Share a few of the failed imaged so I can have a look.
Only when the images are very very poor the solver should fail.
Han,
I just sent you a set of files by WEtransfer. Hope that it reaches you with your sourceforge.net address. If you do not receive anything please let me know.
cs Axel
Han ,
this is the direct link to the files
https://we.tl/t-6jdzCODdlE
Hi Axel.
I will have a look tomorow.
The stars in the files are not round due to your tracking problems. Solving work still fine. For photometry purposes it is better to skip/redo them.
Cheers, Han
Thanks Han, so I will not spend any time more with photometry of the images. Redoing them however will be difficult: the field is central Centaurus and I will not visit the southern hemisphere any time soon.