Resampling Point Clouds Using Series of Local Triangulations

Area of Science:

• Computational geometry
• Computer-aided engineering (CAE)
• 3D scanning and meshless methods

Background:

• Increasing reliance on 3D scanning and meshless methods in CAE simulations necessitates optimized point-cloud geometry representations.
• Existing voxel-based binning methods often compromise geometry and topology, especially with coarse voxelizations.

Purpose of the Study:

• To propose a robust and straightforward algorithm for efficient point cloud upsampling and downsampling.
• To ensure resampling without feature loss or topological distortions, preserving point cloud integrity.
• To provide a method that integrates seamlessly into existing engineering workflows.

Main Methods:

• Development of a Series of Local Triangulations (SOLT) algorithm.
• Utilizing SOLT as an intermediate representation for point clouds.
• Demonstration with mechanically sampled point clouds and real-world 3D scans.

Main Results:

• SOLT enables efficient point cloud resampling while preserving geometry and topology.
• The algorithm avoids complex optimization or machine learning, offering a straightforward approach.
• Resampled point clouds are suitable for solving partial differential equations (PDEs) and surface reconstruction.

Conclusions:

• The SOLT algorithm offers a reliable and high-quality solution for point cloud resampling in CAE.
• This method enhances existing engineering workflows by providing accurate and distortion-free point cloud representations.
• The approach is validated through diverse examples, confirming its practical applicability.