Affiliation
American Association of Variable Star Observers (AAVSO)
Sat, 03/14/2020 - 17:18
I have been having intermittent problems with displaying stacked images in VPhot. The stacked image display appears as all white or all black.
I can go to one of the unstacked sub-images, look in tools>image display, then use those display setting to reset the settings in the stacked image. These new settings will then display the stacked image normally.
I've attached a screen shot of a stacked Ic filtered image with this problem showing my saved comp sequence and the image display settings. If I use Low = 1933 and High = 2717, this stacked image will display normally. The FITS image is in the attached zip file.
The stacked V and B images from this image set display normally.
Phil
File Upload
Phil
Please share images to me at mzk. Could you also note what software you used to collect and reduce before uploading to vphot. Anything new or different than normal? Did you stretch them differently? Bscale or bzero got messed up?
Ken
Ken,
Actually, they aren't blank; they are just scaled very poorly to the point where they look blank white or blank black. The problem stacked image will display fine if I reset the Image Display parameters in the stacked image to match those in one of the unstacked images.
I shared with you the five Ic filtered images of GU Sgr from BSM_S which, when stacked, produce the problem stacked image. The V and B images in this set stacked normally.
I am pretty confident that all the stacking problems I have encountered came from BSM systems, and it is possible that they all came from BSM_S.
A possible clue: When I get BSM_S images they sometimes are listed in the Available Images window as coming from "ACP->Cel.." ( In the header the TELESCOP value is "ACP->Celestron CGEM 5.22".) When this happens, before I do stacking (or any other operations on the images) I rename them in the "Tele" column to BSM_S.
Thanks for looking into this.
Phil
Hi!
I've tried to stack images with ASTAP and encountered similar issues - all white. Any suggestions on which software it should work with?
Magnus
Magnus:
Maxim DL is certainly one of the best imaging software available but it does cost $$! But, I do have more questions first.
1. What software are you using to collect your images?
2. Are you subsequently using ASTAP to both plate-solve and stack images?
3. Do you know that VPhot will both plate-solve and stack fits images? Note that we do appreciate it if users plate-solve before uploading to improve speed through the queue, especially for large images.
Ken
Hi!
I use Ekos/Indi to capture, the AstroimageJ to calibrate and plate solve images. I actually do not use ASTAP for anything for the moment, just tried it for stacking (AstroimageJ do not do that). I do not use Windows for any astro stuff, except producing PEC curves.
My standard software for stacking and processing is otherwise PixInsight, but PI do not play well witjh VPHOT, hence the AstroimageJ process.
I didn't know VPHOT could stack. Is that in the user guide?
Magnus
Yes, it is mentioned in the Guide BUT just take a look at the link tabs on the Image List and play with them. There are not that many. Click on each and see what they say/request.
Ken
Magnus,
Display the images you want to stack (File>Import>Image Sequence). Then on the AIJ toolbar select Image>Stacks>Z Project.
Roy
Ah, thanks!! :) Is that described anywhere? I seem to have missed it in the documentation.
Magnus
I can't remember where I found it. I may heve Googled it.
Roy
Hi again!
Hrm. I tried stacking two images (of BI Lyr), and saved the output as a fits file. Then plate solved it in AstroimageJ, and saved again. THen try to upload it to VPHOT. Well, uploading works, but it does not show up in the list of images. What am I missing?
Magnus
Magnus:
1. Look in "more" (next to Refresh) on Image List. Then look for your target name and system name in the pull down boxes.
2.See if you find your image of interest?
3. IF the new image you uploaded has the identical date/time and filter of the original it would not show up at the top of the image list but at the original date/time order (so older than current).
4. It is probably not a good idea to stack first and then plate solve. Do it in the reverse order. In your case, the stacking probably uses xy pixel position since RA/Dec does not exist. If the images shifted at all the stars would not align properly.
5. Again, just plate solve your original images with AIJ, upload all of them to VPhot, and stack them there.
6. Better still try an experiment, upload two of your original unprocessed images to VPhot. See if they show up in your Image List.
This may be easier to explain on Skype rather than email/post. Email me at kenmenstar@gmail.com.
Ken
Hi!
Thanks! I checked the FITS-header in the AIJ-stacked image, and that was actually quite empty (contained only plate solving info, nothing about the target or exposure time). So I guess that is why it did not show up in the list - not even when trying the More option.
So I did what you suggested and stacked in VPHOT. That worked nicely! So I'll keep doing that and see what happens. Thanks for the offer of Skype - I'll keep that in mind and might take it up, if I run into more issues! :) But remeber, I'm on European time...
Magnus
Hi, this is just a shot in the dark, but here is a theory what is happening:
Is it possible that the tracking for the series of images that caused problems wasn't perfect and there was some drift? When stacking, depending on configuration, this can lead to exceptionally dark stripes at the stacked image's borders (as in your image's top border), which would explain the somewhat strange histogram for your example. So if the histogram gets messed up by the borders that see contributions from only a few frames, this might confuse the scaling mechanism. You can test this by taking the same image, crop out any borders and throw it into VPHOT again. Is the result better?
Again, just a wild guess but I've seen similar stuff happend.
CS
HB
Looking at the header, this is a floating-point image, but with BSCALE=0.18946 and BZERO=7616.47. This is very unusual for a floating-point image. Usually it has BSCALE=1 and BZERO=0 since you can represent a wide dynamic range with exponents. So I think something got screwed up in the stacking. I can display the image just fine in IRAF, but IRAF is VERY flexible when it comes to these things. I see a small amount of miscentering for the stacking, so there is a noticeable stripe on the right hand edge, but the values are pretty normal. Image statistics are min 1408, max 10628 and median 2147, pretty typical.
IIRC, VPHOT is an integer-based system so has limited dynamic range. It has limitations. I bet you ran into one of its FITS reading issues.
Did you do this stack as an average or median? Did you use VPHOT to stack the images, or some other software?
Arne
That stacked image was made from BSM images which came directly to my VPhot account. It was average stacked by VPhot.
Here may be another clue: Around the time this was happening I had been getting some BSM images that were not plate solved. For these images I would use the VPhot "Update WCS" tool. This involved selecting the unsolved images along with a few of the solved images in the same filter. When the tool was applied the unsolved images showed "green" in the WCS column as plate solved. I now think that the image display problems may have occured when I made stacks using any of the WCS updated images.
I haven't had this display problem recently, nor have I recently had any unsolved BSM images in my VPhot account.
I now occurs to me that it may not be necessary for every image in the stack to be solved for the stacking process to work. Next time I get some unsolved BSM images I think I'll experiment stacking them without doing the WCS update.
Phil