This simple handheld device, which is powered by Raspberry Pi Zero, can identify the five most common types of plastic. And it’s the 2021 James Dyson Award ‘Sustainability’ winner.

Congratulations to Jerry de Vos on this impressive prize-winning project! Jerry is from the Netherlands and attended the Delft University of Technology (TU Delft); the scanner was his graduation project as an industrial design engineer. The idea came to him after he worked with Precious Plastic, an organisation that works to reduce waste, and saw how difficult it is to sort plastic for recycling.

Inside the the plastic scanner

Inside the enclosure sits a specially designed breakout board, which emits infrared light towards the object being scanned and measures the light that bounces back; more on this a little further down. A Raspberry Pi Zero feeds the measurement data from that breakout board into a machine learning algorithm, and controls all input and output. The screen tells the user what type of plastic the scanner thinks it has found. The red button starts the scanning process when you press it.

The total cost of all these components is €261.99, with the breakout board taking up nearly €180 of that.

A closer look at the breakout board

When light hits an object, it either passes through the object, gets absorbed, or gets reflected. There are characteristic patterns to the way that different substances — whether that’s different types of plastic, various kinds of paper, or your bathroom tiles — absorb and reflect different wavelengths of light. If we shine light of known wavelengths at an unknown substance and measure what gets reflected back, we can match our observations to the absorbance and reflectance profiles of substances we are familiar with, and this helps us identify our substance. Jerry says:

“The breakout board is the heart of the ecosystem. It shines sequential infrared light on a plastic object, measuring the reflectance at multiple wavelengths. The measurements are communicated over the SPI protocol to an external processor.”

Jerry’s excellent project website features a dedicated section for this breakout board. You can learn where every cent of the €176.97 it cost to build went, and how all the elements work together.

Open source hero

If you would like to build your own plastic scanner, you can find all the details on WikiFactory. There’s a repository with everything you need. Nice one, Jerry.

You can also read Jerry’s full thesis: ‘lastic Identification Anywhere: Development of open-source tools to simplify plastic sorting.