Scholarlink

<p> <?tk 2?>High-fidelity digital twin models of physical entities form the cornerstone of optimized assembly processes. Hull components are characterized by their large scale, substantial weight, and high complexity. Due to the parts manufacturing error and welding deformation, the hull rib plate often gives rise to interference because of the out-of-tolerance of assembly clearance, which has seriously affected the efficiency and quality of ship assembly. Therefore, combining feature point cloud recognition and feature stitching technology, a new fast reconstruction method for large-scale parts is proposed. Firstly, based on the assembly process of parts, an assembly feature library composed of historical data and expert experience is established, and then the constraint features are determined in assembly processing. Training the theoretical assembly model, an improved deep learning network for recognition and segmentation is created. Finally, by utilizing a feature stitching technique that integrates plane clustering segmentation and convex hull algorithm, the large-scale model is reconstructed. Taking a hull block component as an example, the reconstruction speed is improved by 30%.</p>

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

A method for reconstructing large-scale part models based on feature point cloud data

  • 期刊论文

  • 发表日期：

    2025年10月31日

查看全文
  • Jinfeng Liu,
  • Zhilong Cheng,
  • Zhanhu Zhang,
  • Xiaohu Zhang,
  • Yu Chen,
  • Yang Xie,
  • Xiaojun Liu

摘要

High-fidelity digital twin models of physical entities form the cornerstone of optimized assembly processes. Hull components are characterized by their large scale, substantial weight, and high complexity. Due to the parts manufacturing error and welding deformation, the hull rib plate often gives rise to interference because of the out-of-tolerance of assembly clearance, which has seriously affected the efficiency and quality of ship assembly. Therefore, combining feature point cloud recognition and feature stitching technology, a new fast reconstruction method for large-scale parts is proposed. Firstly, based on the assembly process of parts, an assembly feature library composed of historical data and expert experience is established, and then the constraint features are determined in assembly processing. Training the theoretical assembly model, an improved deep learning network for recognition and segmentation is created. Finally, by utilizing a feature stitching technique that integrates plane clustering segmentation and convex hull algorithm, the large-scale model is reconstructed. Taking a hull block component as an example, the reconstruction speed is improved by 30%.