Scholarlink

<p>With the advancement of intelligent manufacturing, the shipbuilding industry is actively pursuing digital transformation in component production lines. However, point cloud data acquired from curved hull plates often contains noise, holes, and missing regions due to specular reflection, occlusion, and a limited field of view (FOV). These data quality issues significantly hinder the application of point cloud-based precision dimensional inspection and process feedback. To address these challenges, this study introduces a novel framework that integrates graph cut with curvature-constrained point cloud processing. Local geometric curvature features are incorporated into the graph cut model for curved plates, improving the accuracy of the energy function in background segmentation. Furthermore, an improved multi-scale curvature-based adaptive filtering algorithm is proposed, which employs iteratively reweighted least squares (IRLS) and radial basis functions (RBFs) to enhance the balance between noise removal and feature preservation. Simulation experiments are conducted using benchmark datasets and state-of-the-art methods for comparative analysis. Compared to existing approaches, the proposed method achieves over 15% higher accuracy in background segmentation. Moreover, the smoothing of data voids reduces surface burrs in plate bending machining, which demonstrates an improved signal-to-noise ratio and increased completeness of the point cloud data. The results confirm that the proposed approach significantly enhances the precision of curved hull plate forming, provides more accurate bending data, and increases the industrial value of digitalization in shipbuilding. The methodology is also applicable to other similar manufacturing domains, indicating broad potential and promising application prospects.</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

Point Cloud Background Separation Algorithm Based on Improved Graph Cuts and Curvature Constraints for Curved Hull Plate Forming

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月2日

查看全文
  • Yimin Zhang,
  • Xijin Zhen,
  • Aimin Xu,
  • Juyan Zheng

摘要

With the advancement of intelligent manufacturing, the shipbuilding industry is actively pursuing digital transformation in component production lines. However, point cloud data acquired from curved hull plates often contains noise, holes, and missing regions due to specular reflection, occlusion, and a limited field of view (FOV). These data quality issues significantly hinder the application of point cloud-based precision dimensional inspection and process feedback. To address these challenges, this study introduces a novel framework that integrates graph cut with curvature-constrained point cloud processing. Local geometric curvature features are incorporated into the graph cut model for curved plates, improving the accuracy of the energy function in background segmentation. Furthermore, an improved multi-scale curvature-based adaptive filtering algorithm is proposed, which employs iteratively reweighted least squares (IRLS) and radial basis functions (RBFs) to enhance the balance between noise removal and feature preservation. Simulation experiments are conducted using benchmark datasets and state-of-the-art methods for comparative analysis. Compared to existing approaches, the proposed method achieves over 15% higher accuracy in background segmentation. Moreover, the smoothing of data voids reduces surface burrs in plate bending machining, which demonstrates an improved signal-to-noise ratio and increased completeness of the point cloud data. The results confirm that the proposed approach significantly enhances the precision of curved hull plate forming, provides more accurate bending data, and increases the industrial value of digitalization in shipbuilding. The methodology is also applicable to other similar manufacturing domains, indicating broad potential and promising application prospects.