Combining different photos of the same objetct
- Thread starter Vulpescu
- Start date
FlapAstro (aka rakla)
Well-known member
fredvanner
Well-known member
Just pick a good registration reference frame and throw the lot at WBPP.
Vulpescu
Member
Vulpescu
Member
Did I got it right, that I have to register all different stacked images and than integrate them?
fredvanner
Well-known member
I assume each group of images will have it's own calibration data. If you just put all the data into WBPP, but with a manually selected registration image (ideally an image with lots of stars) that will define the frame for the combined image, then WBPP will calibrate each group separately, then align them all to your selected reference, and then integrate them all. WBPP has other options, but that should be the default action.
Vulpescu
Member
What do you mean by "it's own calibration data"?
My idea was, not to register and integrate all sub frames, but the final singel stacks.
In my thinking, it should be no further calibration needed, once I deal with allready clibrated, registred and integrated singel pics.
The advice, to use manually a starfull image, is extremly precious, though.
But - by the way - we are talking about the coccon nebula - a star forming region! I supose there are 20 G stars in a field of wiew of 0.3 arc seconds!
My idea was, not to register and integrate all sub frames, but the final singel stacks.
In my thinking, it should be no further calibration needed, once I deal with allready clibrated, registred and integrated singel pics.
The advice, to use manually a starfull image, is extremly precious, though.
But - by the way - we are talking about the coccon nebula - a star forming region! I supose there are 20 G stars in a field of wiew of 0.3 arc seconds!
Last edited:
Vulpescu
Member
Second thougt of mine was to stitch all picks in the stetched state - so to say - when they are allready finallly processed.
Am I wrong with this?
fredvanner
Well-known member
I mean dark frames and flat frames (i.e. each different instrument / camera combination will have separate calibration data). While it is possible to integrate each group separately, then combine the groups, WBPP will probably do better if you combine them all in a single integration (since it will then manage normalisation and scaling on a frame-by-frame basis).
Vulpescu
Member
Got your toughts! Thank you so much - really!
But...:
What do you think about my second post, in which I describe, that I primarily plan to register and integrate the all of the finally procecssed and - therefore - stetched and with individually modificationes made, to individually, look perfect, pics?
My goal is to keep it simpel, if possible. I know, how to use the grooping stuff in WBPP. But I would like to avoid it.
But...:
What do you think about my second post, in which I describe, that I primarily plan to register and integrate the all of the finally procecssed and - therefore - stetched and with individually modificationes made, to individually, look perfect, pics?
My goal is to keep it simpel, if possible. I know, how to use the grooping stuff in WBPP. But I would like to avoid it.
Last edited:
fredvanner
Well-known member
If you try to combine separate processed images, particularly if they have been stretched (i.e. they are no longer linear), you will almost certainly find that they are "incompatible" (i.e. there is no scaling for which their noise characteristics "match"). By all means integrate groups separately if you wish (though, as I note above, it is often better not to), but never perform non-linear transformations (stretches) on separate images before combining them.
["never" is a big word - and there are some exceptions if you are looking for certain types of special effect, but don't do it until you know enough about image processing to know that you want to do it!]
["never" is a big word - and there are some exceptions if you are looking for certain types of special effect, but don't do it until you know enough about image processing to know that you want to do it!]
chris.bailey
Well-known member
With that many data sets I would run each one separately and combine at them afterwardsGot you! Whish, it wouldn't be that complicated. We are talking about 7 or 8 different data sets.
But anyway, thank you so much for your most kind support and advice.
I am more than thankfull...
Take care and - again - thank you,
Christian
fredvanner
Well-known member
Out of interest, why? (I would only run them separately if I had problems running them in a single batch, mainly because a single run allows ImageIntegration to scale and normalise on a frame-by-frame basis across the entire set).
chris.bailey
Well-known member
Largely because 7 or 8 datasets will have widely varying quality (halos, noise, FWHM etc) and importantly orientation. I would rather have 7 different masters to assess and combine than 100‘s of frames. Whenever I have processed a collaborative effort cropping the combination has been tricky, unless every contributor captures at the same rotation angle. I would also be tempted to do DBE, SPCC and maybe even some judicious NR before combining. Sometimes leaving a sub par set out of the combination leads to an improvement.
chris.bailey
Well-known member
Thinking about it further still, these days I would also consider combining starless versions of the dataset masters and then adding the stars back in based on the set with the best stars. It’s a very dense star field so would make sense to me to avoid blobby stars that can result from combining data from different sensor/scope combinations.