Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
I've been messing about with some of my images of Machholz and thought I would share the problems and my solution for stacking images of a moving target
Here is a stack of 17 x 240s @ ISO 800 taken with my Canon 10D. The images have been calibrated with a combined set of 5 darks all 240s @ ISO 800 taken straight after the main image run. The times are 22:43 UT to 23:54 UT 29 Jan 2005. Mean time was 23:15 UT.
The result of stacking these calibrated images aligned on the star field was:
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
Note how the comet has moved and become blurred with loss of tail detail.
OK so you can restack them aligned on the comet nucleus to get this image:
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
Now you can see the tail clear enough - though it is an average of 70+ minutes. However the stars now look like rain drops!
I wondered if I could do better and somehow stack the stars separately from stacking the comet and combine the result at the end.
The first step was isolating the comet. Taking the 17 images aligned on the comet nucleus I combined then using MIN values pixel math - not the MIN/MAX excluded average of the rain drop picture. As the stars don't line up well they are largely removed. Like so:
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
Now you can take this comet-only image and subtract it from each of the original comet aligned images. Then you re-align each of these resultant stars-only images to line up on the stars and stack them to give:
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
Nearly there. The stars-only image is not quite perfect - you can see a residual bit of the comet nucleus. Also the comet-only picture still has stray bits of stars in it. The final image combination step will largely cure these problems. After careful alignment using one of the original 17 images as a template (the 23:15 UT image) I recombined the stars-only with the comet-only image AFTER separately doing Digital Development Processing on them. The final combination step was with MAX values pixel math to get:
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
(sorry about different sizes - making jpegs that fit within forum specs means image detail dictates image dimensions)
This was all done in Images Plus.
Any comments? I actually like the "falling rain" style picture as it contains the most comet detail. But for creating a portrait of both the comet and star field while benefiting from the noise reduction of image stacking this MIN - SUBTRACT - MAX combine method seems to work. I'm just wondering if its a bit too much processed to be valid ?
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Thick_asa_Planck
Dark Sky Hunter
Reged: 09/04/04
Posts: 3341
Loc: UK
|
|
Great tutorial Tonk - i am also in the UK, where abouts are you? I assume you are under dark skies because your Machholz shots are excellent.
Alex
--------------------
It is often commonplace to leave the notation ambiguous - Anonymous
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
Alex, I'm in Headingley Leeds - the photos are taken at a site in the countryside (Armscliff Crag) 15 mins drive to the north. Reasonably dark skies but its a bit of a lovers lane site as well! So I have to set up before 23:00 to deter them - cruel ain't I! The only other good nearby site is Fewston but that is 1 mile from the Men ... with Hill US electronic spy base so I've had 6 police visits/interruptions in the last 6 months - you take your pick! Actually the police haven't been too bad but you end up showing them how to work a telescope  or discussing UFOs  . If I need really dark skies then trips into the Yorkshire dales are possible with a site on Dallow Moor (above Pately Bridge) being very good - but can be rather windswept.
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Corn
Carpal Tunnel
Reged: 12/27/04
Posts: 2198
Loc: Sweden
|
|
Nice work tonk. Looks good too 
--------------------
Canon 300D (mod)
8" f/5 Skywatcher, EQ6 Pro, HEQ5 and Barn door mount.
http://web.telia.com/~u18524382/
|
Anonymous
Unregistered
|
|
Nice work, tonk, and instructive too. Out here in Rural Suffolk, Macholz has been naked eye, fading now though.. about 3.5 to 4 at best.My garden is very dark..lovely . you can see M31 no trouble, naked eye. I even got a nice email from the Don himself when i mentioned the twin nuclei visible telescopically on comet macholz. Great guy, superb Comet finder is he. made my day!at its nearest, you could see it MOVE, obvious really but a first for me.
Thanks
Ps part of the "war on terror" has been the sudden interest by police in anyone Using telescopes round here. There is a lot of USAF , RAF and Nato night flying in East anglia, special forces stuff too.. I went up the lane from my village, next thing I know I was pointing out Starry Night backyard to bored coppers on my laptop.. Beats pub fights etc.
|
Anonymous
Unregistered
|
|
How do you like working with Imagesplus? Is it worth the money?
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
Its worth every penny!
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
I've made a subtle but important improvement to this procedure. The final star-only image is a true average stacking of 17 images of the star field with the minimum values of the comet image substracted from each source image. However the final comet-only image is composed from the minimum values gathered from each source image. This has two problems a) no noise reduction has actually been applied and b) faint tail detail down near the noise threshold will be lost.
I've elimated these problems via the following steps
1) Each of the constituent star-only images (created by subtracting the min-comet-only image from a stars-and-comet image) is further subtracted from the corresponding source stars-and-comet image to create a stack of comet-only images. These are not the same as the min-comet-only image as each one contains a unique noise component ameanable to stacking.
2) This new stack of comet-only images is combined using min/max excluded average (i.e. regular stacking) and then the result is used to create the final image with the stacked stars-only image.
Now we have two correctly stacked and noise reduced stars and comet components in the final image. I've called this the MIN-2SUB-MAX method. The difference between the two methods (the other is MIN-SUB-MAX) is barely noticable but a little more of the comet ion and dust tails and coma is evident:
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Thick_asa_Planck
Dark Sky Hunter
Reged: 09/04/04
Posts: 3341
Loc: UK
|
|
Thanks fore that Tonk - i hope to be going to the Yorkshire Dales with my astronomy club this August, and hopefully, i'll have a barn door and 300D by then, so i can't wait!
Alex
--------------------
It is often commonplace to leave the notation ambiguous - Anonymous
|
Charles
Post Laureate
Reged: 06/12/03
Posts: 4111
Loc: Enterprise, AL
|
|
Excellent images!
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
I've been testing the method out on other comet image stacks. My general conclusion is that it only works well when the imaging conditions have been very good. Firstly a constant sky brightness is required in all the images in the stack as well as a lack of gradients as produced by say light pollution. Also it is essential that image alignment in the stack has to be spot on to avoid artifacts in the subtraction steps.
Here is an example of an image in which a farm building security light added a gradient to the bottom of the image as well as a steady overall brightening of the sky due to the development of thin cloud over the 50 minutes needed to take the stack - the subtraction steps have created odd effects in small parts of the image:
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Chris Graham
mmmm...Haggis
Reged: 04/01/04
Posts: 4870
Loc: Stirling, Scotland
|
|
Good advice and a great image, thanks for posting 
--------------------
-Skywatcher 8" Reflector on HEQ5 with Skyscan
-Orion ED80 Refractor
-70mm Guidescope/grab and go scope
-Canon EOS 350D
-Toucam Pro 2
Astronomy & Veggies
|
Albert van Duin
member
Reged: 01/30/07
Posts: 10
Loc: Beilen, Netherlands
|
|
Thanks Tonk!
I've been experimenting with this subject when I found your very nice tutorial.
One question is still left, you tell you subtracted the comet-only image from each of the stars+comet images in ImagesPlus. How exactly do you do that? Is it a MIN combine?
Thanks in advance,
Albert
--------------------
Albert van Duin
8" F/2.74 ASA astrograph + guidescope + ST4
Modified Canon 350D
(Almost modified) Canon 40D
Losmandy G11 + SkySensor2000
http://www.astropix.nl
|
Tonk
Post Laureate
Reged: 08/19/04
Posts: 4055
Loc: Leeds, UK, 54N
|
|
No its a straight forward subtract using the "Image math ..." menu item
Take a look at this link as it summaries other sometimes better methods - http://tinyurl.com/24dfxy
Finally Deep Sky Stacker (free) written by Luc Coffier (he posts here) is making progress on automating stacking comets. Keep an eye out as at some point soon I'll review DSS's comet handling here (maybe in 3/4 weeks time)
--------------------
Televue 85, GM-8/Gemini, Canon 10D, Canon 450D -DSLRAstromod'ed with Astronomiks clip-in filters (UV/IR, OWB, CLS)
Coronado SM60/BF10, Baader Herschel Wedge
Leeds Sky Clock Ripon Sky Clock
|
Albert van Duin
member
Reged: 01/30/07
Posts: 10
Loc: Beilen, Netherlands
|
|
Hello Tonk,
Thank you very much for your information, I didn't find this thread yet!
Looks like I'll have some reading and testing to do this weekend, after that I can try my hand on my Lovejoy and Linear images:
Lovejoy Linear + M3
Thanks again!
Albert
--------------------
Albert van Duin
8" F/2.74 ASA astrograph + guidescope + ST4
Modified Canon 350D
(Almost modified) Canon 40D
Losmandy G11 + SkySensor2000
http://www.astropix.nl
|
Previous
Index
Next
Threaded
Edit
Reply
Quote
[Re: 