Machine learning breakthrough enables color measurement system for plastics recycling
Researchers at the SKZ Plastics Center have successfully completed an innovative project developing a camera-based measuring system that optimizes color formulation in plastics recycling through machine learning.
A hyperspectral imaging-based color measurement system has been developed to accurately predict chromatic values in recycled materials. (Photo: Luca Hoffmannbeck, SKZ)
The breakthrough demonstrator combines hyperspectral imaging with advanced algorithms to accurately predict color values, with initial tests indicating strong potential for commercial implementation.
The solution emerged through a collaborative project funded by the German Federal Ministry for Economic Affairs and Climate Protection (BMWK) via the Central Innovation Programme for SMEs (ZIM). Working alongside industrial partner inno-spec GmbH, SKZ scientists created a sophisticated color measurement system for the visible wavelength spectrum.
The color measurement system integrates a conveyor belt, specialized LED illumination (upgraded from older halogen technology for improved accuracy), and a hyperspectral camera. By correlating captured color data with readings from commercial spectrophotometers, the team established a robust foundation for machine learning applications, ultimately developing and validating a functional software demonstrator for color prediction.
Rigorous testing at both SKZ and inno-spec facilities validated the system's effectiveness. Researchers used physical test samples and regrind blends to correlate color measurements with imaging data, enabling iterative refinement of algorithms and training of AI models.
Accurate color matching has always been particularly difficult in plastic production, and the growing use of recycled materials has compounded this challenge. While some color variation may be acceptable depending on product applications, manufacturers frequently struggle to predict final colors when working with significant amounts of recycled content.
This uncertainty has driven SKZ Plastics Center to pursue comprehensive research into reliable solutions. The advancement of color measurement system, achieved on April 1, 2025, addresses one of the most persistent challenges in plastic manufacturing.
This pioneering project demonstrates how cutting-edge technologies like hyperspectral imaging and machine learning can transform plastic recycling processes. The successful results not only validate the concept but also create a solid platform for further refinement and eventual commercialization, potentially revolutionizing how the industry approaches color consistency in recycled materials.
"The most precise color formulation possible is crucial for the use of recycled plastic. The successful development of the software in this project was therefore an important contribution to the sustainable use of plastics. The project is a prime example of how important it is to work together on an interdisciplinary basis in a world that is also becoming increasingly complex in technical terms. The broad expertise in a wide range of areas within the SKZ enables us to be a competent development partner, even for complex issues," says Christoph Kugler, Group Manager Digitalization at the SKZ.
