meshwell's avatar
meshwell meshwell@nostr.land 2 months ago
this is what i worked on today for my personal project. this is a satellite image that i requested from mapbox and I used YOLOv11 to detect cars. I downloaded many different images of varying zoom levels from 17.0 - 19.0 binned by 0.5. I determined that 17.5 was the lowest I could go while still classifying successfully. #proofofwork #devstr image

Replies (4)

meshwell's avatar
meshwell meshwell@nostr.land 2 months ago
im gonna make a longer post about it later, but im gonna embark on a long road trip soon and im gonna use this project to find neighborhoods with tons of cars parked on the street so i can blend in and sleep in my car undetected
Fabio Krauss's avatar
Fabio Krauss 2 months ago
Have you thought of turning it into a service? Data as a service. I believe you could sell this service to municipalities, to real estate funds and other investors too
1 replies ↓
meshwell's avatar
meshwell meshwell@nostr.land 2 months ago
i hadnt thought that far ahead but that could be a good idea. im sure knowing the highest clustering of cars on the street could be useful
↑