Frame to frame offset

[Jaap55]

Hi Robin,

For about one year I have been doing deep sky imaging using livestacking.

I live in a Bortle 8/9 area
My equipment is: a 66mm refractor with a FL of 400mm
A QHY715 camera with pixel size of 1,45 micron.
My mount is an old Vixen SXD with Onstep.
With these items I calculate the optimal exposure time with your background calculator. The result is about 8 seconds.
I have a frame to frame offset between 0.1 and 3px.
The average is +/- 1,5px in 8 sec, which is not bad I think

Now I have the following question:
Is it possible to make a filter (like the FWHM filter) to reject the higher frame to frame offset in livestacking?
My end results are quite good, but a filter like this might improve the quality even more.

Thanks for your reply!
Kind regards,
Jaap

[admin]

Hi,

that's an interesting suggestion - I had to think about it for a while, unfortunately I don't think it would always do what you want it to do. Let me give an example where it does the wrong thing...

Capturing frames, and somewhere at a point maybe 75% of the way through a frame, the mount shifts by a few pixels due to poor tracking. Because most of the frame happened before the shift, star positions in that frame will fairly much match the previous frame and the offset will be small - the frame with the shift in it will get used. The next frame is then captured entirely at the new, offset, position. That one will show an offset of the few pixels by which the mount moved and will get rejected based on the offset between the two frames.

Because of that sort of possibility, I think the right thing is to rely on the star sizes/shapes in each captured frame to judge the quality (like the FWHM filter) rather than the offset from the previous frame. There is currently no direct detection of trailed stars, but the FWHM measurement will tend to rise if the stars are stretched out due to trailing. A better detection of trailed stars via the analysis of the shapes of stars is something that I am considering for a future improvement - if I can make something reliable.

cheers,

Robin

[Jaap55]

Hi Robin,

I think your explanation makes sense.
To investigate the influence of the frame to frame offset I did a test last week.I made 60 exposures of 4 sec. each and I saw that a few of them had a high frame to frame offset (higher than 4 px).When viewing and comparing each image I did not see any differences in sizes/shapes between them. So it seems that a higher frame to frame offset does not have a big impact on the final stacked image.
For your information, I make 5 minutes live stackings of images with exposure times of 4-8 sec.As many as possible. These 5 min. stacks are then stacked in Siril to make a final image.I am not using darks or flats, but move the image just a little bit on the laptop screen every 5 or 10 minutes. So that is a kind of manual dithering !
Included is an image of M51 with an integration time of 10 hours.
Regards,

M51-250326.jpg (671.54 KiB) Viewed 2468 times

[admin]

Hi,

that's a nice image of a target that will be tricky from your light polluted observing location - well done

cheers, and clear skies!

Robin