A vision-based system for post-welding quality measurement and defect detection

详细信息    查看全文

• 作者：Hui-Hui Chu ; Zong-Yi Wang
• 关键词：Machine vision ; Image processing algorithm ; Weld quality inspection ; 3D reconstruction
• 刊名：The International Journal of Advanced Manufacturing Technology
• 出版年：2016
• 出版时间：October 2016
• 年：2016
• 卷：86
• 期：9-12
• 页码：3007-3014
• 全文大小：1,590 KB
• 刊物类别：Engineering
• 刊物主题：Industrial and Production Engineering
  Production and Logistics
  Mechanical Engineering
  Computer-Aided Engineering and Design
  
• 出版者：Springer London
• ISSN：1433-3015
• 卷排序：86

文摘

The measurement of the welding bead and the followed evaluation of the weld quality is an important part in the industrial welding. However, the main difficulty in using visual inspection to evaluate the weld quality is the time taken and labor wasted. In order to solve these problems, an automated vision-based system is presented for non-destructive and online post-welding quality inspection. The trouble with respect to vision-based system is the robustness of image processing algorithms. This paper presents a novel image processing method that can automatically extract the weld joint profile and feature points, measure the weld bead size, and detect defects. At the same time, the three-dimensional (3D) profile of the weld surface can be reconstructed with the aim of monitoring the weld quality online. The proposed algorithm is validated through experiment in industrial environment.