GreatAttractor

You're welcome! It will be probably more convenient to use imgalt's successor, ImPPG (also for stacks' post-processing).

That's right, Mesa.

Thanks, I'm glad you liked it.

On 26.11.2016 at 10:32, derekorion said:

Did the original individual frames look much worse than the composite video?

They're the same frames; ImPPG's alignment function only translates and crops the images.

On 26.11.2016 at 10:32, derekorion said:

Also, when the plane is not changing angle in the video, can you make a stacked image of the frames during that period of the video (or from the best original frames from the original video, from more than one period - say if the altered angle then came back to the same orientation a number of seconds later)?

I don't think stacking would be feasible here. Note that the plane's aspect is changing constantly, so stacking would introduce some blur. Also, a plane flying overhead on a (more or less) straight course cannot be visible at the same orientation at all.

It seems to me that for planes you don't really need stacking; rather, a larger aperture and better seeing conditions. See e.g. these cruising altitude photos by dedicated spotters (usually made with 8-10 inch Newtonians and DSLRs): http://forum.rnavspotters.pl/viewtopic.php?f=1&t=1266

On 26.11.2016 at 10:32, derekorion said:

But when slides are prepared and the things being examined are fixed, then I wonder if improved images could be go by stacking.  The only thing I would say is, fixed slides illuminated by a constant bright light, and no changes in the atmosphere, would be that unlike in astronomical imaging where there are moments of good seeing, I guess the opposite would be true of microscope imaging of fixed specimens - lots of good seeing, plus rare moments of poor seeing (e.g. accidental or natural environmental vibrations during imaging).

Sure, you could stack it, but the only effect of stacking static images is increasing the signal-to-noise ratio, not the sharpness. If the specimen is well-lit, this ratio should be good enough already. Perhaps it would help when there's not much light to begin with (fluorescence microscopy maybe?).

Thanks for the library name, looks very interesting. I wrote my one TIFF handling code because once I got annoyed trying to understand libtiff's API.