It is basically an openstreetmap viewer plus a simulator layer that turns the data (https://versatiles.org/) into 3d at a certain level of detail. The Netherlands have very detailed geo data-sets (https://www.pdok.nl/), which are used to augment the 3d representation with stuff like point-cloud rendered buildings (3DBAG), crop-rendering, benches, etc.. Basically, it tries to combine as much public data as possible, to represent a city. The simulator is then used to be able to navigate through it with an avatar and to build basic stuff (Secondlife prim-style + Lua scripting engine).
WebGPU version feels incredibly slow for a Bevy project, crashes constantly, even for the most simplest looking scenes. I'm running it on my high-end desktop
Yes, it is all still very experimental. I think the data pipeline (gathering and constructing the tiles mostly) is still a bit problematic. I mostly test from the Android build at the moment, but I'll see if I can give the wasm build a bit more focus.
Loved it! But I think with advent of llm, there should be a way to somehow feed the current view and generate a realistic 3D view, based on the real data from the same location to be as accurate as possible.
It already uses a harness (lightweight viewer) to generate screenshots with. The agent already sometimes compares it with the aerial layer, but I didn't gave it much thought. Currently, the most difficult things to get right are tunnels and bridges. Comparing it to real location imagery would probably help. I'll give it a try. Thanks for the suggestion!
I like the idea but it seems unusable for me at least at the moment. It crashes whenever I zoom.
I'm curious as what is different between this and say, google maps?
It is basically an openstreetmap viewer plus a simulator layer that turns the data (https://versatiles.org/) into 3d at a certain level of detail. The Netherlands have very detailed geo data-sets (https://www.pdok.nl/), which are used to augment the 3d representation with stuff like point-cloud rendered buildings (3DBAG), crop-rendering, benches, etc.. Basically, it tries to combine as much public data as possible, to represent a city. The simulator is then used to be able to navigate through it with an avatar and to build basic stuff (Secondlife prim-style + Lua scripting engine).
WebGPU version feels incredibly slow for a Bevy project, crashes constantly, even for the most simplest looking scenes. I'm running it on my high-end desktop
Yes, it is all still very experimental. I think the data pipeline (gathering and constructing the tiles mostly) is still a bit problematic. I mostly test from the Android build at the moment, but I'll see if I can give the wasm build a bit more focus.
Loved it! But I think with advent of llm, there should be a way to somehow feed the current view and generate a realistic 3D view, based on the real data from the same location to be as accurate as possible.
It already uses a harness (lightweight viewer) to generate screenshots with. The agent already sometimes compares it with the aerial layer, but I didn't gave it much thought. Currently, the most difficult things to get right are tunnels and bridges. Comparing it to real location imagery would probably help. I'll give it a try. Thanks for the suggestion!