Aperture centroid adjustment

Affiliation
American Association of Variable Star Observers (AAVSO)
Sun, 09/14/2014 - 08:19

I faced great difficulties to adjust the aperture centroid on some targets so to be trustworthy to do time series.

There is a close companion to my target and on a particular image (Img 1) I adjusted the centroid via Pixel/ADU mode, and I saved the sequence. When I try to use that sequence on other images or even I reload the same Img 1 (where the sequenced was adjusted) the centroid is on a different place. That always have happened and does not matter how carefully I have adjusted the centroid and resaved the sequence. When I reload the same image the centroid is on a different place.

Yes I can do the single photometry on every image but they are more than 200. The time series is the easiest way to do this.

Please see the attached screenshots to better understand the situation.

Any suggestions why VPhot mocks me, and how can I vanquish it?

Affiliation
American Association of Variable Star Observers (AAVSO)
Suggestion for Aperture Centroid

Hello PVEA;

Read your post.  I have used VPHOT very little, use Maxim a lot.  Had these kind of difficulties with Maxim and switched from "snap to centroid" to "star matching".  This helped a lot.  Occasionally it misses.  This may not be an option in VPHOT.

I had a thought however.  What if you purposely made the aperture very small, so that it could only capture one star.  Do the locating and centroiding with the small aperture.  Then after all is located, revise the aperture to the appropriate radius of 2-3 FWHM for measuring.  This would hopefully leave the centroid where it was, and would not jump to another star.  

Let us know if this works and if it does not.  As I say, seldom VPHOT, or I would try it myself.  

Gary

Affiliation
American Association of Variable Star Observers (AAVSO)
Aperture Centroid

Hi Gary,

Thank you for your suggestions. As far as I know there is no other options for choosing the centroid except Pixel/ADU mode. It is quite usable to put the centroid to the right place. I used both possibilities:  automatic, with radius 1.5 / 2 FWHM and fixed with different radiuses and initial FWHM. I am using small aperture as 3 px. Before uploading my images to VPhot I always use the star matching (alignment) procedures to be sure that all images are properly aligned.  

It any case the time series photometry works quite well but some of the results are obviously out of range. When I use on that images single photometry the results become quite usable and in the inherent range. So I can do single image photometry to correct only the obviously out of range data.  That means that some of the other results may vary (probably within acceptable uncertainty) but how one can be entirely satisfied with that.

I will continue to search the acceptable time series procedures to make the life easier.

Regards,

Velimir

Affiliation
Astronomical Society of South Australia (ASSAU)
Hi Velimir
I doubt that I can

Hi Velimir

I doubt that I can do better than you but please feel free to share the FITS files via the Share link on the Available Images page and I'll see if I can spot anything you didn't.

David 

Affiliation
American Association of Variable Star Observers (AAVSO)
FITS

Hi David,

A few FITS images are in your folder now. The target is almost in the exact image center. You can spot it using the first screenshot from my first post.

Thank you for the cooperation.

Regards,

Velimir

Affiliation
Astronomical Society of South Australia (ASSAU)
Search radius

Hi Velimir, all

I think I see the problem. See the attached images. The second image (centroid.png):

shows the dialog you get from Tools → Settings. Notice that I have changed the Search radius in pixels to 3, the same as the aperture radius.

Using this setting causes the centroid to be appropriately placed over the target star you identify in your images on the forum, except for one, presumably because it doesn't exist on that image or is too faint. In that image the centroid is over the bright star to the upper left. Making the aperture radius and search radius smaller (not too much!) may help with that.

Once you have comp/check stars, the Time Series dialog could also make use of this search radius value. 

David

Affiliation
American Association of Variable Star Observers (AAVSO)
Aperture Centroid

Hi David,

You are right about the settings of aperture and search radius in the sequence dialog box but the time series does not use them automatically. The time series obviously uses the sequence but not the desired aperture and search radius unless you set them up manually. I have got the good results using your idea and tune up the time series settings. I have used aperture of 3 px, and search radius of 3 px also.

To avoid influence of nearby star I used the parameters for sequence: Aperture=3px, Inner Sky Annulus=10px and Sky Annulus=7.

You can see the attached screenshots with light curves, sequence and time series settings.

Best regards,

Velimir