A feature-extraction localization algorithm research for teaching-free automated robotic welding based on 3D point cloud

The persistent challenge of discrepancies between virtual environments and actual working conditions in offline programming for robotic welding continues to impede practical implementation effectiveness. This research presents a feature-extraction localization technology for teaching-free automated...

Full description

Saved in:
Bibliographic Details
Published in:International journal of advanced manufacturing technology Vol. 138; no. 11; pp. 5397 - 5412
Main Authors: Li, Yiheng, Xu, Yanling, Wang, Xinghua, Ma, Xiaoyang, Wang, Qiang, Zhang, Huajun
Format: Journal Article
Language:English
Published: London Springer London 01.06.2025
Springer Nature B.V
Subjects:
Accuracy
Advanced manufacturing technologies
Algorithms
Automatic welding
Automation
CAE) and Design
Computer-Aided Engineering (CAD
Deep learning
Efficiency
Engineering
Feature extraction
Image registration
Image segmentation
Industrial and Production Engineering
Localization
Manufacturing
Mechanical Engineering
Media Management
Original Article
Quality control
R&D
Registration
Research & development
Robotics
Robots
Simulation
Software
Teaching
Three dimensional models
Virtual environments
Welding
Workpieces
Point cloud registration
Feature extraction
Trajectory correction
Automated programming
Teaching-free welding
3D vision
ISSN:0268-3768, 1433-3015
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The persistent challenge of discrepancies between virtual environments and actual working conditions in offline programming for robotic welding continues to impede practical implementation effectiveness. This research presents a feature-extraction localization technology for teaching-free automated robotic welding based on a 3D vision sensing system. Our approach employs 3D vision sensors to capture the actual workpiece’s point cloud, which is then registered with the standard 3D digital model or used to extract the mathematical model of welds. Subsequently, these registration or extraction outcomes are used to adjust the offline programming trajectory and obtain the precise trajectory on the workpiece. The core technological innovations include the 3D reconstruction of the workpiece, initial localization of point cloud registration through the FPFH-RANSAC-ICP algorithm, and welding seam localization based on point cloud segmentation and feature extraction. Experimental validation on T-pipe, triplanar fillet, and V-groove butt weld configurations demonstrates the algorithm’s efficiency and accuracy.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0268-3768
1433-3015
DOI:10.1007/s00170-025-15797-0