In Panotools::Script 0.29, released just a couple of days ago! There is a tool 
called nona-deshake that does exactly this.

Basically you pick an anchor photo, straighten, crop, defish, or whatever, in 
Hugin, then save the PTO project. nona-deshake then uses this as a reference to 
align all the photos in the set. It even skips already rendered frames so you 
can rerun it to track ongoing projects.

It works well with enormous image sets, though it doesn't cope well with 
day/night transitions, and I had one site camera that slowly tilted to one side 
over a couple of months until it couldn't cope any more.

-- 
Bruno


On 15 May 2019 00:30:13 BST, Bob Campbell wrote:
>So, I’ve had this mad scheme for a while to align a series of
>time-lapse images using a control point finder.  The problem to solve
>is aligning a few hundred images that are very, very close, but due to
>wind or camera movement (it’s a point-n-shoot that re-assesses the
>scene every single shot) the images don’t exactly match up.
>
>My first attempt with align_image_stack blew up since it loads all the
>images at once, and if you throw a few hundred 50 MB tiffs at it, your
>system runs out of memory (well, ok, I don’t have 128 GB in a system
>atm). So another goal is to be able to align an arbitrary number of
>images of arbitrary size, so this can be future-proofed against sensor
>size and my tight wallet. :)
>
>First of all, is there maybe a better tool to use than
>align_image_stack (w/ celeste to clear out clouds)?  I somehow doubt
>the others would handle hundreds of images better, but maybe?  Maybe a
>plugin for Lightroom (or other) that I’m missing?
>
>My train of thought has been:
>- Eat the elephant one bite at a time - do a series of manageable
>stacks, like 3-7 images.
>- Advance to the next stack by one image.  This means that each photo
>is in several .pto files  <- I AM HERE
>- Check for duplicate and very close control points. As luck would have
>it, it does appear to be finding some of the same control points to use
>independently across stacks.
>- Merge common control points and image alignment data into a single
>(huge) .pto file
>- re-run align-image-stack to correct the images.
>
>I’m not quite sure if I’m taking the right next steps. My thought was
>to build a database of control points and keep track of which or how
>many .pto files they’re in, and keep the ones that are in a majority of
>ptos.  Also, since I have multiple correction values for each image,
>should I average the number out or ...?  My intention was that since
>most of the images are stable, the ones that are off would stand out
>like sore thumbs.  
>
>I realize most of the folks here haven’t thought about doing this,
>though a few have since I’ve found other posts via google, but extra
>brains would help! :)  
>
>Slightly off-topic for Hugin, so if there’s a better place to posit
>this, just point me there!

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/C977558C-F9C9-470D-B2FF-61991744ECD0%40postle.net.
For more options, visit https://groups.google.com/d/optout.

Reply via email to