Title

IMPROVEMENT OF SCRAP MANAGEMENT IN EAF AND BOR BY CLASSIFICATION USING COMPUTER VISION ALGORITHMS

IMPROVEMENT OF SCRAP MANAGEMENT IN EAF AND BOR BY CLASSIFICATION USING COMPUTER VISION ALGORITHMS

Authorship

Kempken, Jens

DOI

10.5151/2594-5300-40234

Downloads

heading.download_article_detail 13 Downloads

Abstract

For recycling-based steel production, an on-site scrap quality assessment is mandatory to control the received quality from the supplier. This process can be automated with computer vision technology. We developed an application based on deep learning techniques that assess the scrap by considering different quality characteristics and thus, providing decision support for the scrap yard staff. A graphical user interface allows to continuously review the automatically detected scrap in a simple and quick way. Guided by the scrap yard staff's expertise, the detection accuracy increases over time without specialist support needed from IT experts.

 

For recycling-based steel production, an on-site scrap quality assessment is mandatory to control the received quality from the supplier. This process can be automated with computer vision technology. We developed an application based on deep learning techniques that assess the scrap by considering different quality characteristics and thus, providing decision support for the scrap yard staff. A graphical user interface allows to continuously review the automatically detected scrap in a simple and quick way. Guided by the scrap yard staff's expertise, the detection accuracy increases over time without specialist support needed from IT experts.

Keywords

Scrap management; machine learning; scrap classification; object detection

Scrap management; machine learning; scrap classification; object detection