Could BBC please share the code for matching mountain ridges with locations?
They claim that the first mountain ridge was located by a tip from a Cameroonian source. Here they verify the ridge with the mountain ridges from a height map.
But later on without mentioning any tip or source they "found" an Channel4 news report in archives with again a mountain ridge match. I consider 3) possibilities:
1) They manually watched all the footage from the last few years in that area, freeze framing them whenever they see a mountain ridge, then manually trace it. I don't believe this is what happened, that's too many man hours.
2) They knew through other means (perhaps camera gps location metadata, perhaps report description/summary) that the report video was taken near the massacre video. Because of this metadata/textual data they were able to realize they had nearby footage of an outpost by querying for the massacre location. But then they fail to mention this trivial step and make a bit of a show by highlighting some features and again a mountain ridge. So in this case they have a database of archive footage and filter by location/time.
3) They have automated software for isolating mountain ridge profiles from footage (perhaps edge detection? remove all edges that don't appear to undergo rigid body motion? but how filter clouds etc?), and matching software for locating where on the heightmap it the footage was made. In this case it is explainable without ridiculous man hours or ridiculous showmanship (instead of mentioning say GPS coordinate metadata of the journalism camera). But in this case perhaps BBC could share this code with the public?