Scholarlink

<p>Conventional optical systems for quality inspection often encounter limitations due to the constraints of standard image processing pipelines. The customization required for these systems to function in a production environment is not only laborious and costly, but also lacks versatility for various inspection challenges. In this paper, we introduce an Artificial Intelligence (AI)-powered, adaptable toolbox for Visual Quality Inspection. Our toolbox employs divide-and-conquer methodologies, simplifying intricate tasks into manageable sub-problems that can be addressed with established AI techniques. A user-friendly interface facilitates process monitoring and data collection at the production level, enhancing the AI processing. This innovative strategy promotes the digitization of knowledge via sub-problem annotation, offering a reusable and transferable solution for future Industry 4.0 scenarios. We showcase the efficacy and flexibility of our AI-centric quality inspection approach in different real-world production scenarios.</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

AI-Driven Toolbox for Efficient and Transferable Visual Quality Inspection in Production

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年5月21日

查看全文
  • Patrick Trampert,
  • Sven Mantowsky,
  • Felix Schmidt,
  • Tobias Masiak,
  • Georg Schneider

摘要

Conventional optical systems for quality inspection often encounter limitations due to the constraints of standard image processing pipelines. The customization required for these systems to function in a production environment is not only laborious and costly, but also lacks versatility for various inspection challenges. In this paper, we introduce an Artificial Intelligence (AI)-powered, adaptable toolbox for Visual Quality Inspection. Our toolbox employs divide-and-conquer methodologies, simplifying intricate tasks into manageable sub-problems that can be addressed with established AI techniques. A user-friendly interface facilitates process monitoring and data collection at the production level, enhancing the AI processing. This innovative strategy promotes the digitization of knowledge via sub-problem annotation, offering a reusable and transferable solution for future Industry 4.0 scenarios. We showcase the efficacy and flexibility of our AI-centric quality inspection approach in different real-world production scenarios.