Framework for vision-based robotic corner cleaning of window frames

• Author / Creator

  Tzu-Jan Tung

• Corner cleaning is a critical step after welding of window frames to ensure aesthetic quality. Current methods to clean weld seams are limited in adaptability and quality, increasing rework, cost, and waste. This is largely attributable to the use of CNC cutting machines in combination with manual inspection and seam cleaning. Because the system relies on predefined window designs, cleaning processes are ineffective when dealing with manufacturing imperfections. However, as the blueprint of Industry 4.0 is becoming clear, the development of automation is proceeding at a rapid pace, and new technologies such as robots and sensors are playing a lead role in driving innovation within the manufacturing field. In this thesis, a vision-based robotics system is proposed that enhances adaptability to variability in weld cleaning while ensuring high quality and precision through an approach that combines robotic arms and computer vision in place of existing manual-based methods. The developed system uses edge detection to locate the window and identify its orientation, image segmentation techniques with a pre-trained Mask R-CNN model to detect the window weld seam, and a vision-guided robot manipulator to control the moving path for the robotic arm. The working process begins with the window location to obtain the rough position for the purpose of guiding the robot toward the window target, followed by image processing and detection in conjunction with instance segmentation techniques to segment the target area of the weld seam, and, finally, the generation of cleaning paths for further robot manipulation. The proposed robotic system is validated in a simulated environment as well as in a real-world scenario, with the results obtained demonstrating the effectiveness and adaptability of the proposed system.

• Subjects / Keywords

  • machine vision
  • window manufacturing
  • vision-guided robotics
  • deep learning

• Graduation date

  Spring 2022

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-n0kd-q913

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.