Jerry de Vos' Plastic Scanner Turns a Raspberry Pi Into a Spectrometer for Plastics Recycling
Designed for sorting plastic by type, the handheld scanner runs local machine learning analysis on a Raspberry Pi Zero W.
Graduate industrial design engineer Jerry de Vos has turned a Raspberry Pi into a low-cost handheld scanner, which identifies plastic types to ease sorting for recycling — and has taken his place among the winners of the 2021 James Dyson Award.
"I got the idea while working with Precious Plastic and seeing firsthand the difficulty in sorting plastic types for recycling," de Vos explains of the project, the Plastic Scanner. "Looking for possible solutions I stumbled on the open source research project for the ReReMeter. My project set out to implement the discrete infrared spectroscopy technology of the ReReMeter into a product that can be used anywhere and is easy to build, user-friendly, and widely available."
Driving the handheld version of the Plastic Scanner is a Raspberry Pi Zero W — the more powerful Raspberry Pi Zero 2 W, with a quad-core processor, having not been released at the time of development — with a custom-designed breakout board connected via SPI.
"The breakout board shines eight individual LEDs with a wavelength of 850, 950, 1050, 1200, 1300, 1450, 1550 and 1650nm," de Vos explains. "The LEDs have a narrow bandwidth with a full-width half max of 20nm and a tolerance of just 3nm. The reflection from these LED lights can be measured by the sensitive and precise analog- to -digital converter, the ADS1256."
The Plastic Scanner, then, is effectively a device for infrared spectroscopy — but using a discrete approach that makes it considerably cheaper than traditional lab equipment. "This method is slightly less accurate," de Vos admits, "but can still identify the most common plastics.
"After the user has turned on the handheld scanner, the screen tells them to push the scanning button once a scan is desired. As soon as the Raspberry Pi recognizes that the button is pressed it sends a command to the breakout board, which starts the scanning procedure. Once the scanning is completed it is communicated back to the Raspberry Pi, which subsequently starts the interpretation algorithm to interpret the sample. Once this is completed the result is shown on the screen."
De Vos has released details of the project via Wikifactory under the reciprocal GNU General Public License 3.0 with more available on GitHub, but notes that the final documentation has yet to be uploaded. More information is available on the project website.
For his efforts, he has been awarded the 2021 James Dyson Prize in Sustainability — joining projects in other categories including a wearable glove sensor designed to measure eye pressure and evaluate glaucoma risk and a handheld tool for quickly treating stab wounds.
More details on all the winners are available on the James Dyson Award website.