Revolutionizing robotized assembly for wire harness: A 3D vision-based method for multiple wire-branch detection
- Authors
- Nguyen, Thong Phi; Kim, Donghyung; Lim, Hyun-Kyo; Yoon, Jonghun
- Issue Date
- Feb-2024
- Publisher
- Elsevier BV
- Keywords
- Automation system; Convolutional neural network; Machine vision; Wire harness assembly
- Citation
- Journal of Manufacturing Systems, v.72, pp 360 - 372
- Pages
- 13
- Indexed
- SCIE
- Journal Title
- Journal of Manufacturing Systems
- Volume
- 72
- Start Page
- 360
- End Page
- 372
- URI
- https://scholarworks.bwise.kr/erica/handle/2021.sw.erica/119438
- DOI
- 10.1016/j.jmsy.2023.12.002
- ISSN
- 0278-6125
1878-6642
- Abstract
- Automating stages for deformable objects in the production line, such as assembling a wire harness into a predefined position, presents a complex task due to the specialized characteristics of these objects. This complexity is further magnified when dealing with a wire harness that has multiple branches. Designing the robot motions and performing the assembly tasks for such target objects not only requires the 3D wire profile of each branch but also entails the identification of each branch. To satisfy this requirement, this study introduces a novel systematic method that processes the point cloud data of the wire harness, captured from a high-quality 3D camera, which provides the constructed wire profiles of branches in the captured frames. The developed method is a combination of a technique to classify the colour of wires and a detecting-matching method to identify the wire terminals, which provides a convenient approach for the proposed tracking method to define the 3D wire profile. The validation of the proposed approach in a real-world robot system not only highlights its incredible usefulness but also demonstrates its immense practicality. This ground-breaking method has the potential to revolutionize factory environments by replacing human labour, leading to substantial cost reductions and increased efficiency.
- Files in This Item
-
Go to Link
- Appears in
Collections - COLLEGE OF ENGINEERING SCIENCES > DEPARTMENT OF MECHANICAL ENGINEERING > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.