Scholarlink

This paper presents a method for inspecting the quality of the rocker arm surface. The rocker arm is manufactured using the hot forging method, which often results in a less smooth surface due to scaling and the potential for surface imperfections. To automate the inspection of the rocker arm surface, a deep learning approach was applied to detect scratches. The proposed method begins by capturing high-resolution images of the rocker arm. These images are then divided into multiple sub-images, and the YOLOv4 algorithm is applied to detect scratches on the surface. The developed method was tested on a real dataset and achieved a precision of up to 96.4% for detecting scratches.

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

Rocker Arm Surface Inspection Employing Deep Learning Technique

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Thanh-Hung Nguyen,
  • Van-Huy Tran

摘要

This paper presents a method for inspecting the quality of the rocker arm surface. The rocker arm is manufactured using the hot forging method, which often results in a less smooth surface due to scaling and the potential for surface imperfections. To automate the inspection of the rocker arm surface, a deep learning approach was applied to detect scratches. The proposed method begins by capturing high-resolution images of the rocker arm. These images are then divided into multiple sub-images, and the YOLOv4 algorithm is applied to detect scratches on the surface. The developed method was tested on a real dataset and achieved a precision of up to 96.4% for detecting scratches.