Computer vision can be an important part of machine learning applications at a variety of scales, from $20,000 Tesla robots and self-driving cars to smart doorbells and vacuum cleaners. It’s also a challenging task because, compared to a cloud infrastructure, “real” edge devices typically have much more limited hardware specifications.
YOLO (You Only Look Once) is a very popular object detection library, its first release was made in 2015. YOLO is particularly interesting for embedded devices because it can run virtually anywhere – not only Python, but also C++ (ONNX and OpenVINO) and Rust versions are available. A year ago, I tested YOLO v8 on a Raspberry Pi 4. Nowadays, a lot has changed – a new Raspberry Pi 5 and a newer YOLO v10 were released. So I expect a new model on new hardware to run faster and more accurately.
The code presented in this article is cross-platform, so readers who don't have a Raspberry Pi can also run it on a computer running Windows, Linux, or OS x.
Without further ado, let's see how it works!
Raspberry Pi
For someone who may have never heard of Raspberry Pi, let’s do a quick recap…
