Hey everyone,

After working on the Plastic Scanner for quite some time, I wanted to give you all an honest update on where things stand.

What we’ve accomplished

We’ve made some real progress on this journey. We built several working prototypes and this year we got them evaluated by the photonics department at The Hague University of Applied Sciences. The results are published in a paper, and showed we could correctly identify about 70% of plastic samples, which is pretty cool, but not quite where we need to be for a reliable tool.

Current challenges

Unfortunately, I’ve hit a few roadblocks that are tough to overcome right now:

Accuracy: 70% isn’t reliable enough for practical use
Parts sourcing: Key components are difficult to find and get consistently (source)
Time: I simply don’t have the bandwidth to give this project the attention it deserves

Because of these challenges, I’ve decided to pause active development on the Plastic Scanner for now. This wasn’t an easy decision, but I think it’s the right one given the current limitations.

The project lives on

Here’s the thing: all the documentation, designs, and code are still online and open source. If you’re excited about plastic identification technology and if you know how to take it further, please feel free to build on what’s here! Maybe you’ll figure out the accuracy puzzle or find better component sources. If you get it to work, please let me know and keep it open source 😉

Thank you

Huge thanks to everyone who helped make this project possible. You can see the full list of contributors on our About page. But seriously your support, feedback, and encouragement meant the world to me.

What’s next?

I don’t know, I might return to this project down the road when circumstances are different, but for now I’m stepping back.

Thanks for being part of this journey!

All project files, documentation, and designs remain available in the repository for anyone interested in continuing development. (source, source)

Hi, we are building a device that can identify the most common types of plastic to make plastic recycling simpler, more accessible, and more viable! All in an open source way. We don’t know yet what it is going to look like, but maybe something like these:

DIY Version of the Plastic Scanner
Visual model from thesis Jerry

Commercial Version of the Plastic Scanner Visual model by AED student team

Current status

We’re currently in the prototyping phase and working towards a version that is recommended to build or purchase. Stay informed with our monthly progress updates through our video series. Watch the latest update now here:

Why we are doing this

As you might know, plastic recycling is important, but putting this into practice can be quite challenging. One of the big things is identifying and sorting plastic.

Plastic pollution is a well-known problem worldwide and is still growing. It negatively affects humans and wildlife through animal death, groundwater pollution, and the incorporation of microplastics in our digestive system. There are many initiatives focusing on reducing the negative effects of plastic pollution, but the amount of plastic consumed and the subsequent pollution is still increasing every year. Additionally, during the COVID-19 pandemic, the dependency on single-use plastics has increased exponentially.

That is why it is important to keep improving recycling infrastructure, especially in low and middle-income countries. Their plastic waste management is often informal, and tools are insufficient for the correct management of plastic waste, resulting in plastic pollution. The research conducted in this thesis showed that especially the sorting stage of the plastic recycling process is very time-consuming and labor-intensive. This discovery led to the central research question: which resources can be developed to accelerate the process of plastic sorting for informal recyclers?

Discrete near-infrared spectroscopy makes it possible to identify over 75% of all plastic used in everyday life. Therefore, it became my mission to make this technology accessible to recyclers in low and middle-income countries.

By applying principles of context variation, local manufacturing, and open development in the design process, tools that accelerate the sorting of plastic waste were created. This brings Plastic Identification Anywhere another step closer, with the end goal of fighting plastic pollution, together, today.

read the full thesis

Endeavors

The goal of this project is to have more plastic recycled by the end of the day. We want to craft the tools that are needed to make the work of recyclers anywhere around the world easier, simpler, and more accessible. To get there we have defined endeavors, small specific projects with a fixed start and end point for people to help out with their skills.

? Coding- Compare datasets

?Taken by: Hendrik ?
We have scanned quite different plastics with multiple spectrometers. what can we learn from comparing these different datapoints?

➡️Prerequisites: does not require any previous endeavor to be done
⏱️Duration: 5 afternoons
?Expected outcome: A visualization of the good and bad of each dataset, and hopefully some interesting insights.

? Coding- Write Machine learning model

Based on the data we can get some 3D scatter plots, but the goal is to have a computer do the interpretation.

➡️Prerequisites: Does not require a specific endeavor to be ready
⏱️Duration: 2 weeks
?Expected outcome: a pre-trained model with explanation, scikit learn or tensorflow

? coding- determine data flow/storage

The size of the measurement data is quite low, but ultimately we would like to function the device fully offline, with the possibility to update a pretrained model once in a while

➡️Prerequisites: Does not require a specific endeavor to be ready
⏱️Duration: 2 weeks
?Expected outcome: a example program/setup with dummy data

? coding- Squeeze the bug

We sometimes run into situations where we don’t know whats happening. Help us what is wrong with our python programming

➡️Prerequisites: we need to have buggy code
⏱️Duration: ongoing
?Expected outcome: code that works 😉

? video- Capture the spirit

Throughout the year we are going to work on some cool projects, it would be cool to capture this by video or photo

➡️Prerequisites: can be done at any time
⏱️Duration: a couple of days
?Expected outcome: Videos or photos of the project

? Integration – build the measurement setup

To analyze if software improvements work, or if hardware modifications have a good outcome it is good to have a “standardized” platform to do a variety of tests.

➡️Prerequisites: some prior hardware prototyping skills
⏱️Duration: 3 weeks
?Expected outcome: something like an optical breadboard to setup various experiments

? integration- Build minimal viable product

Build a prototype to capture plastic scans in the field, something that is battery-powered with wifi connectivity to grow our dataset.

➡️Prerequisites: Some prior hardware prototyping skills
⏱️Duration: 4 weeks
?Expected outcome: A device with an interface for input and a screen for output, to capture data

✅? Spectroscopy – characterize optics

Date finished: June 2023
Work by: Queena
Resource: https://docs.plasticscanner.com/research

✅? Spectroscopy – Most important LEDs

Date finished: June 2023
Work by: Queena
Resource: https://docs.plasticscanner.com/research

✅? coding- Streamline PS plot

Date finished: March 2023
Work by: Joost
Resource: https://github.com/Plastic-Scanner/PSplot

✅? Spectroscopy – determine Optimal distance between sample and sensor

Date finished: March 2023
Work by: Peter
Resource: Drive link

✅ engineering – PCB for handheld scanner

Date finished: December 2022
Work by: Markus, Gustav, Lawrence
Resource: https://github.com/Plastic-Scanner/handheld-scanner

✅ Research – Article about NIR spectroscopy

Date finished: October 2022
Work by: Jerry
Resource: https://plasticscanner.com/working-principle/

✅ Spectroscopy – Reference material set

Date finished: December 2022
Work by: Peter
Resource: Standard reference kit of plastic available

✅ engineering – find ideal backing material

Date finished: November 2022
Work by: Nemo & Jerry
Resource: Sintered PTFE sheet from Thorlabs

✅ engineering – Firmware for DB2.1

Date finished: April 2022
Work by: Jure
Resource: https://github.com/Plastic-Scanner/DB2.x-Firmware

✅ engineering – Software for DB2.1

Date finished: June 2022
Work by: Jure and Joost
Resource: https://github.com/Plastic-Scanner/PSplot

Open source

One of the core values of the Plastic Scanner Project is the open source mentality. We see this as an essential part of the project that helps to get the right feedback and guidance from experts, as well as reach the people that need a plastic scanner the most.

At the heart of this open source project are the developers, the people who actually make it possible, the ones who chip in with their skills, the people who make it to the about us page, so find out here who is helping out making the Plastic Scanner project a reality!

Next to this is also the extensive documentation about the working principle of the Plastic Scanner. Documentation about the project, background information, and build instruction can be read here:

Read the documentation

Come and help

We are always looking for bright minds that want to help out making plastic identification possible. Especially data science, spectroscopy, and electronics. If you want to chip in with your skills please read the page here:

Work at the Plastic Scanner

Want to help out but don’t have the skills or time at this moment to be able to help out hands-on? You can also help out the project tremendously by doing a (one-time) donation, for this, check out the options here:

Donate to the plastic scanner

FAQ

Can I buy it?

No, at this point we are fully in development mode, so testing a lot of things and also changing the design rather quickly. This is why we do not sell anything (yet) and why we also do not recommend it at this point to build the development board for other intents as being an active developer.

Can my company help to supply plastic?

For testing the plastic scanner in different environments, we already have a big list of alpha and beta testers.

Can my company help with the development of the project?

probably yes! if you can help us analyze samples or build prototypes then we are super happy with your help!

How does this whole open source work?

We have a specific license that says you can replicate the Plastic Scanner and even commercialize it, but if you make any improvements you are required to keep that open source and share it back with the community.

Join the updates

Join our once-in-a-month email list and

Stay up to date