I don’t want to hear about your Plex, your NPM, your notes application or science forbid, your budgeting application. I want to hear the most exotic thing you setup to selfhost, that probably only you and a hand full of people around the world actually use or even need. A problem that you solved in a way, that makes people go WTF. Go!

I’ll start: I live in the mountains, and there is snow, lots of snow. I often tell people “We had 3m of snow last year”, but is that really true? So, I thought to myself: Can you measure snowfall? It seems you can, so I setup a USH-9 ultra sound measuring device, connected it via IC2 to my Home Assistant and now I can tell people with confidence, that we had a total of 3.45m of snowfall last season, with max snow height of 60cm on January 5th.

Future project: I have chickens. They lay eggs. I have cameras. I want to know which hen lays how many eggs. Solution? AI image recognition of the hens (who is who) and if they have laid an egg. Any inputs welcome.

  • OutsidePerception911@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    7 months ago

    I’m up for the chicken challenge, dad wants to have chickens, we live too far apart. It’s reusable, plus now you can know which chickens will go to nuggets and which give you omelette

    • ElevenNotes@alien.topOPB
      link
      fedilink
      English
      arrow-up
      1
      ·
      7 months ago

      Problem is training the model which hen is which from different angels. I would need to provide a lot of video material for every single chicken and then apply ML to get a match. With 40 birds, that’s a lot of prime video footage per hen. Maybe I’m missing a better solution?