For our final project we’ll try something a bit more complicated. We’re going to leverage the extra grunt of the Pi 4 (this will work on a Pi 3 but it won’t be fun) and the TensorFlow machine learning software to enable the Pi, via a camera, to classify (or attempt to classify) whatever we show it. We used the High Quality camera module, but the basic one will work fine. We’ll be using the TensorFlow Lite package for Python, which enables lower-powered devices (or anything not equipped with the latest tensor processing chiplets) to run pre-trained models without relying on the cloud.
It’s easy to connect a camera to your Pi. utility mentioned earlier.