Geometric Modelling for 3D Point Clouds of Elbow Joints in Piping Systems

Area of Science:

• Geoinformatics
• Mechanical Engineering
• Computer Vision

Background:

• Pipe elbow joints are critical in various systems, but their spatial data extraction from point clouds is challenging.
• Existing methods lack a universal geometric model for diverse elbow joint types.

Purpose of the Study:

• To propose a novel geometric model for analyzing pipe elbow joints using 3D point cloud data.
• To adapt the model for common elbow types (90° and 45°) and demonstrate its utility in applications.

Main Methods:

• Development of a geometric model with translational, rotational, and dimensional parameters.
• Utilizing simulated and real laser scanning point cloud datasets for model verification.
• Application of the model for point cloud registration and mounting bracket detection.

Main Results:

• Accurate geometric recovery of simulated joints with minimal translational (0.3 mm) and rotational (0.064°) errors.
• Real-world data fitting achieved 97.2% accuracy in diameter estimation.
• Joint-based registration accuracy reached sub-decimeter and sub-degree levels.

Conclusions:

• The proposed geometric model effectively analyzes pipe elbow joint geometry from point clouds.
• The model facilitates applications like precise point cloud registration and component detection.
• This approach enhances the spatial analysis capabilities for piping systems.