Combining Star Aligned and Comet Aligned images

[MikeWiles]

Hi,

I shot a few frames of Comet ISON last night and it looks pretty good.  I've aligned using star alignment and integrated the image.  I have a nice, starry background with a mangled comet from the data rejection as expected. 

Then I used the Comet Alignment module to align on the comet and create an integrated image and the comet looks great.  The background?  Not so much.  Does anyone have some tips on how to intelligently combine these two integrated frames so that I can keep my comet and the nice background too?  I know it's a basic question, but I've no idea really where to start.

Thanks in advance for any help you can offer.

Mike

[Ignacio]

I did this for comet PanSTARRS and Lemmon, early this year:

http://www.pampaskies.com/gallery3/Comets-Asteroids/PanStarrs
http://www.pampaskies.com/gallery3/Comets-Asteroids/Lemmon_montaje

What I did is use the comet-aligned version (stars rejected using median combine), on top of a star-aligned stacked subset of frames , where the comet move just a little (used 3 frames for PanSTARRS, and a single frame for Lemmon, above). This depends on how fast the comet moves with respect to the stars, how long your subs are, and if there is wait time in between subs.

One other tip: if in your comet-aligned sequence, the stars are too close together for effective rejection, you may want to integrate 2 or three different subsets, where you skip on/two subs in each sequence. Then, you combine the 2 or 3 stacks into one. I did this for Lemmon (still not super clean).

Hope it helps,
Ignacio

[pfile]

i think if you get 2 clean images, you can probably use some kind of pixel math operation to merge the two images.

you could do max(stars,starless) for instance to pick the maximum value from both images, creating a new image.

or i suppose you could make a star mask, apply that star mask to the starless image, and then using the pixel math expression "stars" apply pixel math to the starless image.

or maybe some combination, like iif(stars>threshold, stars, starless) applied to the starless image, which would replace the data in the starless image only if the data in the starry image is brighter than some threshold value. maybe thru a mask or not, you'd have to experiment.

rob