Researchers in Norway are building a machine learning model to better understand where plastic waste comes from and to predict where it will accumulate.
The model is in the form of an open source app that people can use to take photos of plastic waste from a specific region and find out where it comes from. Machine learning can then determine which trends exist in terms of the type of plastic that is being handled in regions (e.g. plastic waste in agriculture or industrial areas) and predict which areas will be hot spots and where for plastic pollution.
“The ultimate goal is to be able to influence politics with the data,” says PhD student Christina Hellevik.
Christina Hellevik is one of the researchers at the Norwegian University of Science and Technology (NTNU) leading the pilot project, aptly named PlastOpol. The aim of the project is to structure the cleanup process and develop a database that can be used by local authorities.
Building a database through photos
According to Hellevik, most of the beach cleanings in their area are carried out by volunteers who are tasked with collecting the litter and then manually registering the plastic via a portal. Besides being slow, she notes that it can demoralize volunteers because the process is cumbersome. That way, volunteers can quickly take photos of plastic as they clean, and the app can recognize the plastic and automate the registration process in a database.
The database allows the pollution trends to be visualized, while the model can provide clues as to where targeted cleaners should be organized and even indicate who the main polluters are.
“One of the ultimate goals is to make this cleanup more professional and to shift the burden on the authorities as well,” says Hellevik.
This means that precise polluters will be held more accountable through government intervention and political decisions. It also gives the government the responsibility to organize more “professional” cleaning teams for handling plastic waste.
In cooperation with the port authority of the Ålesund region, the district of Møre og Romsdal and the municipality of Ålesund, PlastOpol is a pilot project with the aim of wider use. Volunteers will test the app this spring and it should be ready for public use by 2022.
“We hope that in a larger project we can combine the predictive model with the app and create a comprehensive tool for people all over the world,” says Hellevik. “But that could take a few more years.”
Read more about a new way of separating microplastics from wastewater.