One of the challenges of recycling plastics is to determine what type of plastic a given item is actually made of. In order to meet the demand, large recycling centers rely on various automatic systems to separate different types of plastics from the incoming stream. But in less technologically advanced areas of the world, workers will find themselves having to manually identify plastic objects; This is a time-consuming and error prone process.
To improve this situation, [Jerry de Vos], [Armin straller] and [jure Vidmar] have been developing a handheld open hardware device, which they call a plastic scanner. They hope that their pocket device can accurately identify various types of plastics by measuring the reflectivity of plastics to infrared light in the field. The device is guaranteed to be very easy to operate because the user only needs to put the device close to a piece of plastic, press the button, and then wait for the information to pop up on the OLED display.
Or at least, it’s an idea. Although the team eventually hopes to release a kit to build its own handheld plastic scanner, it seems that the hardware is not fully ready for production. The latest work seems to have been put into the development board, which allows the team to improve their processes, which is not surprising. The development unit combines an infrared LED array with wavelengths ranging from 850 to 1650 nm, an InGaAs photodiode connected to an ADS1256 24 bit analog-to-digital converter (ADC) and Arduino UNO. In contrast, the final hardware uses a raspberry skin zero and a smaller “breakthrough board”, which contains sensors and infrared LEDs.
Browsing the software library of the project, we can see that the device uses python, tensorflow lite and the infrared reflectance value database of known plastics to try to determine the closest match. Obviously, the accuracy of such a system depends largely on the amount of known good data, but at least for now, it seems that users have the responsibility to establish their own collection or IR values.
Although this project is interesting, we have some doubts about its pure optical method of identifying plastics. The automatic recycling center does use an infrared spectrometer, but this is only one of many tools used. Without additional data points, such as the density or electrostatic properties of the plastic being measured, the plastic scanner seems to have a fairly high error range. Only considering the various textures and colors that users may encounter when using the device will be a considerable challenge.