Scholarlink

<p>This research compares the performance of the latest state-of-the-art YOLO (You Only Look Once) models, specifically YOLOv5n, YOLOv6n, YOLOv7-Tiny, and YOLOv8n. For this purpose, images of cardboard packaging with various damaged and undamaged samples were utilized. The classification and identification process considered five categories: undamaged, face damage, edge damage, corner damage, and top damage. Experimental results were obtained for each of the four models, reporting average precision, average recall, <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="11760_2025_4088_Article_IEq1.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="18" /> </InlineMediaObject> <EquationSource Format="TEX">\(F_1\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>F</mi> <mn>1</mn> </msub> </math></EquationSource> </InlineEquation>-score, and mAP@50, which assess the accuracy level in categorizing damage to the packaging. This study aims to provide a solution using computer vision capable of performing more effective classifications to support individuals conducting visual inspection processes based on attributes and mitigate workplace blindness. Thus, it reduces costs in reverse logistics and indirectly benefits the reduction of environmental pollution caused by various transport methods used in reverse logistics.</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

Performance evaluation of YOLO models for damage identification in tertiary packaging

  • 期刊论文

  • 发表日期：

    2025年4月16日

查看全文
  • Daniel Arturo Olivares-Vera,
  • Emmanuel Ovalle-Magallanes,
  • José I. Hernández-Vázquez,
  • José O. Hernández-Vázquez,
  • David A. Gutierrez-Hernandez,
  • Angela del Pilar Olivares-Vera

摘要

This research compares the performance of the latest state-of-the-art YOLO (You Only Look Once) models, specifically YOLOv5n, YOLOv6n, YOLOv7-Tiny, and YOLOv8n. For this purpose, images of cardboard packaging with various damaged and undamaged samples were utilized. The classification and identification process considered five categories: undamaged, face damage, edge damage, corner damage, and top damage. Experimental results were obtained for each of the four models, reporting average precision, average recall, \(F_1\) F 1 -score, and mAP@50, which assess the accuracy level in categorizing damage to the packaging. This study aims to provide a solution using computer vision capable of performing more effective classifications to support individuals conducting visual inspection processes based on attributes and mitigate workplace blindness. Thus, it reduces costs in reverse logistics and indirectly benefits the reduction of environmental pollution caused by various transport methods used in reverse logistics.