Scrap is today a primary raw material pushed by the decarbonization objectives and the growth of EAF-based electric transition of steelmaking. Such a shift is worldwide increasing the scrap use. On the other hand, scrap quality strongly impacts product properties and process efficiency. Therefore, tools allowing use of low-quality scrap via detection and reduction of contaminants, e.g. copper and steriles, are needed. This contribution presents a Deep Learning-based solution to automatically identify copper impurities from images of scrap on a moving conveyor belt, which is developed within the European project PURESCRAP. Results from field tests are presented and discussed.
A Deep Learning-Based Approach to Recognize Contaminants in Scrap for Improved Scrap Quality
Colla V.
;Vannucci M.;Siddique A.;Petrucciani A.
2026-01-01
Abstract
Scrap is today a primary raw material pushed by the decarbonization objectives and the growth of EAF-based electric transition of steelmaking. Such a shift is worldwide increasing the scrap use. On the other hand, scrap quality strongly impacts product properties and process efficiency. Therefore, tools allowing use of low-quality scrap via detection and reduction of contaminants, e.g. copper and steriles, are needed. This contribution presents a Deep Learning-based solution to automatically identify copper impurities from images of scrap on a moving conveyor belt, which is developed within the European project PURESCRAP. Results from field tests are presented and discussed.| File | Dimensione | Formato | |
|---|---|---|---|
|
paper_PURESCRAP_published.pdf
non disponibili
Tipologia:
PDF Editoriale
Licenza:
Copyright dell'editore
Dimensione
1.3 MB
Formato
Adobe PDF
|
1.3 MB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
