Raspberry Pi Zero 2 W Detects Heat Anomalies with Machine Learning

By Ash Hill published about 18 hours ago

Adding the “smart” to your heating system via the Raspberry Pi.

(Image credit: Rob Lauer)

Building the DIY smart home of your dreams is now a step easier thanks to the Raspberry Pi and the ever-growing market of sensors. The many electronics and maker communities improve the build process by providing seemingly endless tutorials and creations to follow along with like this thermal anomaly detection project created by Rob Lauer at Hackster.

The project monitors Lauer’s home heating system, using AI to keep an eye out for unexpected changes in temperature. This isn’t his first Pi project either, we recently covered a Pi-powered speed trap created by Lauer which uses a Raspberry Pi to detect when passing cars have exceeded the speed limit.

In a nutshell, the Pi monitors a boiler in Lauer’s basement using a thermal camera module. A custom machine learning model was created by Lauer to train the Pi to recognize familiar/safe thermal readouts from the system. The data acquired by the Pi is transmitted to a cloud service using a wireless cellular module for the Pi.

Sponsored Links

Best of free games to play nowHero Wars

For the main board, Lauer opted to use the new Raspberry Pi Zero 2 W board paired up with a Blues Wireless Notecard Raspberry Pi kit which provides wireless IoT functions including a cellular connection and GPS capability. A Pimoroni MLX90640 thermal camera is used to capture thermal images from the heating system.

The ML model for anomalous heat detection was created using Edge Impulse. In this use case, Lauer is monitoring his boiler but it could easily be adjusted to monitor any other heat-generating hardware. The Blues Wireless Notecard isn’t critical but it provides some extra functionality that might be worth the investment depending on what you want to monitor and where it’s physically located.

If you want to create this project yourself or just get a closer peek at how it all goes together, check out the thermal anomaly detection project page over at Hackster. Lauer has provided everything you need to get started including a full parts list and tutorial for setting up the software side.

To dig into more inspiring Pi projects, visit Rob Lauer’s profile at Hackster for additional projects and be sure to explore our list of best Raspberry Pi projects for more incredible creations from the maker community.