The molten state of scrap within aluminum furnaces is still inspected manually in the majority of aluminum production plants. This paper presents an automated solution for on-line scrap monitoring, notifying the readiness of the furnace to process new batches of scrap. A Computer Vision based deep learning model has been developed to process the continuous video acquired by a camera located inside the furnace. The model can identify the presence or absence of scrap in the melting furnace and notify the end of the melting process. Evaluation of the model across various image resolutions demonstrated a high accuracy of up to 99%. Tests performed on an industrial plant demonstrated the validity of the proposed solution.
Scrap Monitoring in Aluminum Melting Furnace using Computer Vision and Deep Learning
Ravi Y. S.;Marelli S.;Chiariotti P.;Tarabini M.
2024-01-01
Abstract
The molten state of scrap within aluminum furnaces is still inspected manually in the majority of aluminum production plants. This paper presents an automated solution for on-line scrap monitoring, notifying the readiness of the furnace to process new batches of scrap. A Computer Vision based deep learning model has been developed to process the continuous video acquired by a camera located inside the furnace. The model can identify the presence or absence of scrap in the melting furnace and notify the end of the melting process. Evaluation of the model across various image resolutions demonstrated a high accuracy of up to 99%. Tests performed on an industrial plant demonstrated the validity of the proposed solution.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
