Recent PostsRevisiting the Triangulum Galaxy Deep Sky Astrophotography with your Camera and Lens Part 2 Deep Sky Astrophotography with your Camera and Lens Part 1 A Golden Horeshoe Collaboration: M31 Andromeda Galaxy IC 1805: Gnarly Creatures In Space The Power of Image Stacking Comet Processing for Non Trailing Stars and Comet Focal Reducer test: Celestron SCT f6.3 and 80mm refractor (2008 article) Adventures in Astrophotography: How I got started (2008 article) Welcome to Weather and Sky Photography's Blog
Comet Processing for Non Trailing Stars and Comet
Comet Hartley October 8, 2010
When processing a long exposure multiple sub frame shot of a fast moving comet you are usually faced with the decision of either stacking and aligning the sub frames on the comet and getting trailing stars or stacking/aligning on the stars and getting a trailing comet. The free software Deep sky stacker (DSS) has a star freeze method that will attempt to give you non trailing stars and comet but it may not always work out. So in this case you are faced with a processing challenge. Below I briefly outline my method. The Comet Hartley image comprised of 29x 3min subframes totalling 1hr and 27min. with the QHY-8 and Astro-Tech 8in RC. To start the process, you will have to create two stacked images with DSS.
This is a result of getting DSS to stack on the stars using sigma clip. Basically it is stacked as though it's a typical astro image ignoring the fact that a comet is in the field. The stack is then stretched and colour corrected in Photoshop.
This is the result of a DSS stack that was run using the comet stacking mode and Sigma Clip option. In DSS you may have to manually locate the comet in each frame. The resulting stack will show a comet that is sharp and a background field that is blurred with obvious banding from star trails. The stack is then stretched and colour corrected in Photoshop. Side note: 3 min sub frames was too long for this focal length so the core of the comet was elongated. I minimized it's appearance by using the "shift darken" technique.
Cleaning up Stack 1 and 2
Now Stack 1 and 2 will be combined and processed together but both images need some further cleaning up. After processing Stack 1 you are left with a green fuzzy patch. To get rid of that simply run it through Pixinsight using the dynamic background extraction tool (or use whatever method you prefer) in order to get rid of gradients and the remnants of the comet. The result of it is seen above. Then clean up Stack 2 by applying noels horizontal or vertical banding removal action to remove the blurred star trails. If the trails or banding happens to not be vertical or horizontal simply rotate to either orientation and run the appropriate banding removal action. After running the action and the background is free of banding or trails, rotate it back to the original orientation when done.
Combined and Processed Stack 1 and 2
To combine Stack 1 and 2 simply add them together in screen mode in Photoshop. Providing you make sure they are in their original orientation and didn't crop the image they should line up accurately. If not, overlay a reference frame to nudge the comet frame into the accurate position. Finally create a new layer with the merged stacks and apply your finishing touches to the image.
" In DSS you may have to manually locate the comet in each frame"
Actually DSS has a largely undocumented feature where alignment for comets uses the image time-stamps to improve the accuracy of the comet alignment. You only need to select the comet in the first and last image of the sequence - and if neither of these are the reference frame - then mark the comet in the reference frame as well. DSS uses the time stamp in the data header of each RAW or FITS image.
This feature has been in DSS since around 2007 after I asked Luc to try out the method for me - it just appears to have never been officially documented. One noticeable feature is that if you (say) do an average combine then you will see that the star trails are perfectly linear and are free of any minor wiggles that follow the small errors that manually marking each frame introduces.
I hope this is useful for you. :)
No comments posted.