I could see this as especially useful for climbing, hiking, archeology and other activities. This reminds me of photosynth, mixed in with photogrammetry and other technologies. When images are taken they are geo-tagged but another layer is added, by detecting the angle the picture was taken at. Whether it was looking down, up or in another direction. In other words in three, rather than two dimensions. This is great for modelling, but also for helping people get a sense of the images that they’re looking at. Climbing pictures, where we can see the upward or downward angle would be interesting.
It would also be fun to write a blog post, find the api information, and include that within a geojson file, to provide location info, angle, and other data. This data could then be ported from a photo app, to a hiking app, to a website and more. With more images 3d models become possible. Imagine gathering Via ferrata images and documenting the route.
Of course they can also use this for mapping glacier progress, or regression, coastal erosion and a multitude of other topics. I found this talk really interesting, and I see plenty of uses for such data.