Scholarlink

With the advancement of deep learning, industries are now demanding higher standards for assembly product quality and efficiency. This study integrates temporal action localization, human pose estimation, and graph comparison algorithms, utilizing a multi-thread concurrency mechanism to facilitate real-time monitoring and assessment of assembly behavior. Such an approach enables monitoring personnel to focus solely on videos depicting abnormal behavior, thereby enhancing the overall quality and efficiency of the assembly process. In Sect. 2, we conduct an overview of factory assembly behavior techniques. Subsequently, we introduce the technology roadmap integrated into our system, present the architecture, demonstrate experimental results on our dataset, and draw conclusions.

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

A Learning-Based Monitoring System for Factory Assembly Behavior

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yi Liu,
  • Hao Wang,
  • Xiaodong Liu,
  • Man Jiao,
  • Wenchang Li,
  • Xiaochuan Wang,
  • Ruijun Liu

摘要

With the advancement of deep learning, industries are now demanding higher standards for assembly product quality and efficiency. This study integrates temporal action localization, human pose estimation, and graph comparison algorithms, utilizing a multi-thread concurrency mechanism to facilitate real-time monitoring and assessment of assembly behavior. Such an approach enables monitoring personnel to focus solely on videos depicting abnormal behavior, thereby enhancing the overall quality and efficiency of the assembly process. In Sect. 2, we conduct an overview of factory assembly behavior techniques. Subsequently, we introduce the technology roadmap integrated into our system, present the architecture, demonstrate experimental results on our dataset, and draw conclusions.