Deep learning teams up with robotics

6. February 2020 // All topics, News,

EVK, the image processing specialist, and the Swedish company OP Teknik have jointly developed a solution for the efficient sorting of construction and demolition waste. By combining hyperspectral imaging technology with a deep learning approach from EVK together with robot technology from OP Teknik, a flexible system has been created for high material throughput with high sorting accuracy.

By combining EVK Hyperspectral Imaging Technology with a deep learning approach and robotics technology from OP Teknik, the two companies developed a flexible, high- throughput system SELMA for accurate sorting of selected fractions in C&D waste. The system combines EVK HELIOS imaging platform with robotics-based sorting and is characterized by the ability to accurately differentiate between many heterogeneous materials, such as paper, cartonnage, wood or even light bulbs. The stand-alone sorting system SELMA can be rapidly deployed to remote demolition sites and addresses recyclers to whom high C&D sorting accuracy and high throughput is of high importance.

Dr. Matthias Kerschhaggl, Chief Technology Officer at EVK, explains: “Providers of sorting systems like OP Teknik benefit from a company like EVK because it offers not only imaging sensing technology but also analytics software and domain relevant application knowhow. The EVK HELIOS platform, for instance, helps machine builders like OP Teknik in developing their solutions within shortest time, thus giving them a time to market advantage”.

Fredrik Brandt, CTO, OP Teknik explains: “The classification enables the handling of complex material flows, as they occur in CnD waste sorting”. He adds: “The provided sorting system SELMA can be used anywhere and is ideally suited for large sorting plants which handle industrial waste”.

Alexander Fetz
+43 316 461 664

Share this page: