What kinds of image processing are used for this? Do the images need to be aligned, or can telescopes be pointed precisely enough? Are the images combined using mean/median, or something more sophisticated than that? What settings are the original photos captured with?
There are several software options on image processing for astrophotography. Do a search for 'astrophotography image stacking', and you'll get a list of software, tutorials, videos, etc. A couple of the popular ones are Deep Sky Stacker[0] or PixInsight[1] or even Photoshop. They offer different options/capabilities.
The main thing about the capture settings is to use RAW. Other settings ISO/exposure time/etc is dependent on camera being used. However, whatever you can do to capture as much light as possible within each frame is the goal.
The software does image alignment rotate/scale/etc to do the stacking. You can stack images taken of the same object from different physical locations. Spend a weekend in the desert shooting an object, then spend another weekend the next month at the top of a mountain shooting the same object, and all of the images can be stacked.
The telescope alignment precision is important, but less so than it used to be for a couple of reasons. With gear available today, you can take "portable" telescopes into the field, do a decent polar alignment and then allow the guide scope/software to correct for any imprecision of the main scope's alignment and even tracking issues from manufacturing issues with the mount's worm gear. A guide scope is a second smaller telescope (wider field of view) attached to the main scope with a camera attached to it. That camera is connected to a computer running the guide software, and will track a designated star. The guide software will talk to the telescope's motors, and can speed up/slow down the motors to keep the guide star to within a 1/4 pixel deviation.
Also, with digital cameras, images of shorter exposure times are taken and then stacked in software. There's multiple benefits to doing this. Consider exposing a single frame for 60 minutes, or 12 5 minute exposures, or 30 2 minute exposures. If anything bad happens during that exposure (a plane or a satellite crosses your view, someone uses a laser pointer through your frame of view, a bug lands on your primary, etc) it's not "that big of a deal" to capture it again. Also, digital camera sensors tend to get noisy with longer exposures due to heat build up around the sensor (a problem film cameras do not suffer).
> Also, digital camera sensors tend to get noisy with longer exposures due to heat build up around the sensor (a problem film cameras do not suffer).
Maybe worth pointing out that film has its own issues with long exposures, though. If I remember right, film's response to light isn't strictly linear with exposure time so you get less and less useful additional exposure as you expose longer.
Yes, with film there is the Schwarzschild-effect which causes the sensitivity and the color reaction vary with the exposure time.
While digital camera sensors usually pick up noise for long-time exposures, this is less an issue for astronomical cameras, because they fight this noise by cooling the CCD-chip. Usually the chip is cooled via a Peltier-element to temperatures below -20C, where thermal noise is very low.
According to the official website [1] (which is linked to from TFA), the images were processed using PixInsight, a program popular among amateur astronomers. This page: [2] explains how PixInsight performs image alignment; it turns out to be a pretty complex (and interesting) process. The same page also explains the process of merging images.
