What does this research mean for the field?

A three-part processing pipeline combining short-wave infrared hyperspectral imaging, 3D convolutional neural networks, and mixed integer optimization can characterize steel scrap composition with approximately 76% accuracy to optimize recycling process parameters. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

This work aims to accurately determine scrap composition to enhance recycling process parameters in steel production.

June 20, 2026Waste Management & Research The Journal for a Sustainable Circular Economy0 citationsOpen Access

Hyperspectral Imaging and Deep Learning for Steel Scrap Composition Optimization

Hyperspectral scrap characterisation for scrap composition optimisation in steel recycling

View Full Paper

Ask AI

Bookmark

Authors

HGHeimo Gursch AOAndreas Ofner RHRobert Harb

Discussion

Member takes

Overview

Randomized trial investigates scrap composition optimization in steel recycling, suggesting improved recycling efficiency.

Key Points

This work aims to accurately determine scrap composition to enhance recycling process parameters in steel production.
Implemented a three-part processing pipeline including hyperspectral imaging, deep learning, and optimisation.
Utilized 437 spectral bands in the short-wave infrared range for hyperspectral imaging.
Compared the performance of 2D CNN, 3D CNN, and multilayer perceptron, with 3D CNN showing the highest accuracy.
Achieved approximately 76% accuracy in detecting 14 material classes from scrap.
The 3D-CNN outperformed other models in identifying material classes.
Optimisation focused on minimizing energy use and additives while adhering to storage constraints.

Cite This Study

Gursch et al. (2026) studied this question.

synapsesocial.com/papers/6a362ee4db0793dc1a5368d9 https://doi.org/10.1177/0734242x261451604

View Full Paper

Ask AI

Bookmark

June 20, 2026Waste Management & Research The Journal for a Sustainable Circular Economy0 citationsOpen Access

Hyperspectral Imaging and Deep Learning for Steel Scrap Composition Optimization

Hyperspectral scrap characterisation for scrap composition optimisation in steel recycling

View Full Paper

Ask AI

Bookmark

Authors

HGHeimo Gursch AOAndreas Ofner RHRobert Harb

Discussion

Member takes

Overview

Randomized trial investigates scrap composition optimization in steel recycling, suggesting improved recycling efficiency.

Key Points

This work aims to accurately determine scrap composition to enhance recycling process parameters in steel production.
Implemented a three-part processing pipeline including hyperspectral imaging, deep learning, and optimisation.
Utilized 437 spectral bands in the short-wave infrared range for hyperspectral imaging.
Compared the performance of 2D CNN, 3D CNN, and multilayer perceptron, with 3D CNN showing the highest accuracy.
Achieved approximately 76% accuracy in detecting 14 material classes from scrap.
The 3D-CNN outperformed other models in identifying material classes.
Optimisation focused on minimizing energy use and additives while adhering to storage constraints.

Cite This Study

Gursch et al. (2026) studied this question.

synapsesocial.com/papers/6a362ee4db0793dc1a5368d9 https://doi.org/10.1177/0734242x261451604

View Full Paper

Ask AI

Bookmark