Alignment worked quite well with a selection in the center.
Then I stacked the whole thing and after the background extraction the image is well balanced and contrasted, but there’s literally nothing coming out of the stacking ! 120 images stacked are like one image in terms of details.
I’ve already got some great results with the milky way at 24mm, and used the same process here.
Any idea why ?
Thanks for your help !
Hello, it may be difficult to reliably identify stars with such a small sampling, and if the sequence is long enough, some will appear, other disappear, and the horizon circle will have to move, leading to a bad result. Warping will also make things look bad.
Maybe you can share the first and last image you want to stack so we can see what’s happening?
Unless you meant to do so, sky should be on the outside. Even so, you will likely lose some star detail if you do so, so having the sky as the planet sort of makes sense.
Hey thanks for your answers here, really appreciated !
Here are the first and the last image of the series.
So, @vinvin you’re right, the horizon is moving but most of the sky is the same from first to final image.
The stack won’t work for parts that are close to the horizon, but the aim was to have this large portion of sky stacked, at least (aka see plenty of objects at once), and then see if I could re-composite the final with the foreground.
And, the alignment seemed to work well !
But really I can’t see any difference between single frames and stacked image, so that’s the weird part !
@lylejk Sorry but I didn’t catch what you meant there
Sorry no, it’s not about that at all !
It’s kind of the opposite, it’s a domemaster format, or fisheye, it’s meant to be seen in a dome, like a planetarium !
could you show a snapshot of the options you are using for stacking?
I suspect that, if you are using input normalization, well, your images are not well suited for the underlying assumptions we have. “Typical” astrophotography images have mostly sky while your images have let’s say 50/50 unrelated-dark background and the rest of sky (+very bright patches on the horizon changing from frame to frame). This will for sure mess up the normalization step.
But let’s see first if that’ve the case and then, we’ll try to devise smthg to find a workaround.
Hey Cecile, sorry it took me some time to plunge back into the project.
So, I’m not sure what you wanted to see, but here’s snapshots of what I used (I think) for alignment and stacking.
It’s pretty much defaults settings, with global alignment and homography, based on a selection at the center of the image (as large as I could), then mean stacking with pixel rejection and additive with scaling normalization.
Nothing wrong as such, but maybe not well suited for your images. Could you just try doing the stacking with “no normalization” instead of “additive with scaling”. Just to see if the statistics are messing things up?
So I tried again with no normalization and I couldn’t see any difference between the 2 results.
I’ll post a snapshot just to show you what it looks like after stacking, BG extraction, chromatic calibration and CLAHE. This is what I get with, or without normalization.
So obviously there’s plenty of annoying things on the image because of the foreground moving, but this is really the first test I managed to make in 360 imagery stacking, hopefully next time I’ll find clear space all around.
The thing is I don’t find it interesting in terms of what comes up in the sky thanks to the stacking, compared to what I usually get with 24mm shots of the milky way…
What do you think ?
