I wonder how this compares to purely vision-based systems which use nothing but the images themselves for stabilization. Here are some quite old results of stabilization using image-based 3d-reconstruction of the scene which I wrote more than 10years ago, compared with other stabilization programs of that time (Deshaker, Adobe After Effects, Youtube). With todays improved hardware and progress in 3d-algorithms you may not need any additional gyroscopic data.
"Viewsheds" of any location can be calculated and matched with photographs using "GeoImageViewer", an application I wrote a couple of years ago. Any feature in the photo can be interactively identified in a mapview and vice versa, including the boundary of the viewshed. See the link below for a couple of examples.
I wonder how atmospheric refraction is handled in the calculations for the longest line of sight. Since it (a) strongly affects the line of sight, and (b) depends on temperature and weather, how is a static "world record" possible, or even defined? E.g. objects may appear 400m higher in 200km distance under typical conditions.
Atmospheric refraction is due to the vertical gradients of atmospheric pressure, temperature, and composition of the atmosphere, all of which are usually not precisely known, and which vary with time, so one gets larger lines of sights at certain times. For my application I used the standard formulas for astronomical refraction (-> many weblinks) with constant medium gradients. If I recall correctly this results in ~100m height correction for features in 100km distance and ~400m in 200km distance (features appear higher than without atmospheric refraction). For your application it would make sense to use two extreme values for the gradients to get maximum and minimum, as you suggested.
"Viewsheds" of any location can be calculated and matched with photographs using "GeoImageViewer", an application I wrote a couple of years ago. Any feature in the image can be interactively identified in a mapview and vice versa, including the boundary of the viewshed. As has been mentioned in the comments, it is essential to include atmospheric refraction in the calculation, at least for distances above ~100km.
What is definitely wrong in this article is the claim, that only bears eat living prey. Also wolves do that. There are many cases where wolves attacked animals (cows, sheep, deer), and as soon as those are immobilized, start feeding. Sometimes these animals survive with horrible wounds. Many images and videos can be found via google.
Combined cycle power plants take heat at 2000°C using a combination of gas and steam turbines. Their conversion efficiency to electric power is 52%, see Wikipedia. This is proven technology, at least 30 years old, and quite a few of those exist out there.
Steam turbines alone can be operated at higher temperatures by using mercury instead of water. Some plants using this approach were built in the 1920/30s.
I mentioned GeoImageViewer in a reply above. It contains several algorithms to determine the location of a landscape photo given some control points selected on a map. It is also able to determine lens parameters (fov, distortions,...).
https://hdersch.github.io/
GeoImageViewer, not an overlay but side-by-side view of photograph and maps. Clicking the image anywhere (not just precalculated pois) shows corresponding location in map and vice-versa.
https://hdersch.github.io/
Nernst sold the patent in 1902 for 1 Million Mark to AEG, which corresponds to roughly 10 Million EUR today. It was one of the highest valued patents at the time, and is still quite impressive today.
A comparison with one of the many SIMD-mathlibraries would have been fairer than with plain libm. Long time ago I wrote such a dual-platform library for the PS3 (cell-processor) and x86 architecture (outdated, but still available here [1]). Depending on how standard libm implements atan2f, a speedup of 3x to 15x is achieved, without sacrifying accuracy.
There are many projects displaying georeferenced photographs using Leaflet, which usually work via precalculated hotspots in image- and mapview. Here's a project that uses photogrammetry to link not just some but all points. Clicking the photograph zooms into the corresponding position in the Leaflet-view and vice-versa.
https://www.youtube.com/watch?v=-m3fwhx3Z5g