✈️ Pi-lane Spotting

About Pilane

Phone setup

I live quite close to an airport, I can see planes flying overhead all the time. I was on the terrace one day with a couple friends and was quite surprised by the number of planes that flew over us. It was a Friday and thus, Pi-lane was born. It was a joy to build this. You can find a more detailed write-up here.

This project is heavily inspired by skybot.cam (RIP), but tries to be a bit more DIY without using an ADS-B receiver or a Cloud vision API.

How it works:

  • The camera is an old Android Nokia device. I pull the images from it using adb and scrcpy
  • The images are copied to a Raspberry Pi running YOLOv11
  • The Pythons script uses the Flightradar API fetches the details about the aircraft and the flight

You can find the script here.

It ends up missing a few flights since inference on the Pi 3B+ takes about 2.5s per image, even when using the NCNN export of YOLOv11. I could upgrade the Pi but using an underpowered device is part of the fun, I really enjoyed the challenge.

Wishlist:

  • Show some nice stats on the website (all images of a callsign, of an airline etc.)
  • Purchase a RTL-SDR and use that instead of an API (completely self hosted!)
  • Identify fewer birds as planes 🐦
    • Yolo bird
There's also a Telegram channel where the script posts the images: https://t.me/pilanespotting

Thank you for visiting Pilane!