Built as the capstone for the UIUC Internet of Things course, this system lets a Raspberry Pi run a small greenhouse by itself. The Pi reads a temperature and humidity sensor, watches the plants through a camera, and turns a fan, heater, and humidifier on or off to keep the climate steady.

The same camera images were used to train a TensorFlow model that recognizes whether a crop has sprouted, and the Pi runs that model on each new photo it takes.

A companion iOS app, written in Swift, shows the current temperature, humidity, and latest photo, reports what the sprout model found, and lets you control the fan, heater, and humidifier from anywhere. The app and the Pi stay in sync through Google Firebase.