I did pretty much the same thing some 10 years ago but only for my own images and this creator overcame a problem that I could not. I ingested a collection of images, read the metadata and used that to place the images on a map like this. However, I could not figure out how to determine facing. This project has an arrow indicator to show this, I ended up setting it manually. I would love to know how they determined this because I really cannot figure it. Does image metadata contain more information than it used to? Is there a way to use an AI agent to analyze the light/angles/streets/landmarks to approximate this?
3 comments
[ 2.8 ms ] story [ 20.1 ms ] threadBut why is it in Russian, even when EN is active?
Can be compensated somewhat by chosing a different OSM-layer, though.