<p>The mechanical properties of aluminum alloys vary significantly across different grades due to differences in their chemical compositions. Therefore, when aluminium scraps are properly sorted during recycling, they can be more easily returned to their specific applications and yield a considerably higher value compared to unsorted scraps. Using computer vision for aluminum scrap sorting could potentially enhance the efficiency of the recycling process. In this work, we test and verify whether computer vision-based methods can be successfully deployed for commercial aluminum scrap grade classification using images only, on both the pre-shredding and post-shredding scraps. Image datasets of both pre-shredding and post-shredding scraps are carefully curated. The experimental results and the feature analysis show that the computer vision-based methods can classify the post-shredding scraps with high accuracy (<InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40831_2025_1172_Article_IEq1.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="43" /> </InlineMediaObject> <EquationSource Format="TEX">\(98.4\%\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>98.4</mn> <mo>%</mo> </mrow> </math></EquationSource> </InlineEquation>) using the traces left by the shredding process, showing the potential of applying computer vision-based classification for accurate commercial aluminum sorting.</p> Graphical Abstract <p></p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Computer Vision-Based Aluminum Scrap Grade Classification and Detection for Upcycling

  • Yijun Quan,
  • Matthew Dunn,
  • Giovanni Montana,
  • Zushu Li

摘要

The mechanical properties of aluminum alloys vary significantly across different grades due to differences in their chemical compositions. Therefore, when aluminium scraps are properly sorted during recycling, they can be more easily returned to their specific applications and yield a considerably higher value compared to unsorted scraps. Using computer vision for aluminum scrap sorting could potentially enhance the efficiency of the recycling process. In this work, we test and verify whether computer vision-based methods can be successfully deployed for commercial aluminum scrap grade classification using images only, on both the pre-shredding and post-shredding scraps. Image datasets of both pre-shredding and post-shredding scraps are carefully curated. The experimental results and the feature analysis show that the computer vision-based methods can classify the post-shredding scraps with high accuracy ( \(98.4\%\) 98.4 % ) using the traces left by the shredding process, showing the potential of applying computer vision-based classification for accurate commercial aluminum sorting.

Graphical Abstract