CertAIn - AI for improved waste flow statistics and higher sorting rates at recylingcenters

Purpose and goal

The project aims to develop AI-based computer vision models to continuously and automatically identify and thus generate data on what type of objects visitors leave in load carriers for different fractions at recycling centers. With such data, it is possible to steer away from today´s manual picking analyses, which are time-consuming and, due to high labor costs, are only done sporadically. Being able to digitize the picking analyzes is something that many municipalities are asking for, and with the help of more data, a good foundation is created to reach higher sorting rates.

Expected effects and result

With the help of more data about such waste that is left at recycling centers, and which is also continuously generated, a valuable basis is created for municipalities to be able to more accurately develop and test development strategies and efforts that can improve waste sorting. Increased sorting rates in turn provide good conditions for achieving increased resource efficiency through improved material recycling. In the slightly longer term, an opportunity is also seen in being able to use the AI models to develop a type of interactive sorting aid for visitors.

Planned approach and implementation

The project lasts for 18 months and is divided into six work packages which together include hardware installations, data collection, creation of "datasets", development of AI models, full-scale tests and dissemination of results. Chalmers Industriteknik is responsible for and leads the technical development and where the municipalities play a key role when it comes to data collection, setting requirements and verification of results. The data collection will be done at two recycling centers in Gothenburg and Kungsbacka.