Scholarlink

<p>This study presents a consistent mathematical framework to compress a densely populated 3D Point Cloud Data (PCD) set into a low-resolution subset of “picked” points by employing the Peridynamic Differential Operator (PDDO). This subset provides sufficient accuracy for recovering the missing “unpicked” points. An initial subset of "picked" points is randomly selected and subsequently the “picked” set of points is increased adaptively by considering the deviation from the original data. The recovery of the original PCD is achieved by considering a volumetric structured grid which entirely embeds the “picked” (compressed) data set. The robustness and capability of this approach are demonstrated by considering synthetically generated challenging data with sharp corners and edges. Subsequently, it is applied to benchmark data representing a bunny and a dragon. Finally, it is applied to the real LiDAR data of a terrain. It has proven to be a robust approach for compression and recovery of PCD in the presence of irregularities, edges, corners and complex geometry.</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

Peridynamic differential operator for lossy compression and recovery of 3D point cloud data

  • 期刊论文

  • 发表日期：

    2025年9月16日

查看全文
  • Erdogan Madenci,
  • Atila Barut

摘要

This study presents a consistent mathematical framework to compress a densely populated 3D Point Cloud Data (PCD) set into a low-resolution subset of “picked” points by employing the Peridynamic Differential Operator (PDDO). This subset provides sufficient accuracy for recovering the missing “unpicked” points. An initial subset of "picked" points is randomly selected and subsequently the “picked” set of points is increased adaptively by considering the deviation from the original data. The recovery of the original PCD is achieved by considering a volumetric structured grid which entirely embeds the “picked” (compressed) data set. The robustness and capability of this approach are demonstrated by considering synthetically generated challenging data with sharp corners and edges. Subsequently, it is applied to benchmark data representing a bunny and a dragon. Finally, it is applied to the real LiDAR data of a terrain. It has proven to be a robust approach for compression and recovery of PCD in the presence of irregularities, edges, corners and complex geometry.