Scholarlink

<p>Due to the effect of aging machines, injection molded machines (IMMs) with configured parameters would produce defective IMPs. Thus, after obtaining the information of injection molded parts (IMPs) defects, the parameters of IMMs would be optimized to produce qualified IMPs. To limit the computing power of IMMs and random defects of IMPs, we propose a <b>double</b>-<b>p</b>hase <b>d</b>etection <b>a</b>lgorithm (Double-PDA) for IMPs defects, including IMPs key-frame extraction module and improved YOLOv5n module. On one hand, the Key-FEM can filter useless and repetitive frames of the video stream, reducing unnecessary defect detection time. On the other hand, im-YOLOv5n detects small target defects of IMPs, which has lightweight parameters and acceptable feature extraction ability. Extractive experiments demonstrate that the proposed method can effectively achieve real-time defect detection of IMPs under hardware conditions of Intel Core i5-7400 CPU.</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

Simplifying complexity: a double-phase detection algorithm for defects of injection molded parts within the limited computer source

  • 期刊论文

  • 发表日期：

    2025年8月21日

查看全文
  • Xie Wei,
  • Wu Haorui,
  • Liang Haoming,
  • Zhang Langwen,
  • Yu Xiaoyuan

摘要

Due to the effect of aging machines, injection molded machines (IMMs) with configured parameters would produce defective IMPs. Thus, after obtaining the information of injection molded parts (IMPs) defects, the parameters of IMMs would be optimized to produce qualified IMPs. To limit the computing power of IMMs and random defects of IMPs, we propose a double-phase detection algorithm (Double-PDA) for IMPs defects, including IMPs key-frame extraction module and improved YOLOv5n module. On one hand, the Key-FEM can filter useless and repetitive frames of the video stream, reducing unnecessary defect detection time. On the other hand, im-YOLOv5n detects small target defects of IMPs, which has lightweight parameters and acceptable feature extraction ability. Extractive experiments demonstrate that the proposed method can effectively achieve real-time defect detection of IMPs under hardware conditions of Intel Core i5-7400 CPU.