Point cloud denoising method for annular forgings based on dual-path feature fusion.

Area of Science:

• Engineering
• Computer Science
• Materials Science

Background:

• Accurate dimensional measurement of annular forgings is vital for quality and safety.
• Environmental factors and manufacturing issues create noisy point clouds, hindering precise measurements.
• Existing methods struggle with conflicting global and local geometric data.

Purpose of the Study:

• To develop a novel point cloud denoising framework for accurate measurement of annular forgings.
• To address challenges posed by non-uniform point clouds and noise interference.
• To improve the precision and reliability of geometric measurements in industrial settings.

Main Methods:

• Proposed a dual-path structured denoiser (DSD) framework.
• Implemented a collaborative global-local modeling approach.
• Integrated a priori geometric constraints and confidence-based assessment.
• Utilized a multi-layer refinement structure for progressive outlier filtering.

Main Results:

• DSD effectively extracts local topological structures and global geometric patterns.
• The framework demonstrates superior noise suppression capabilities.
• Experimental results show DSD outperforms state-of-the-art methods in precision, recall, and F1 score.
• High shape preservation was achieved on both industrial and benchmark datasets.

Conclusions:

• The dual-path structured denoiser (DSD) framework offers a robust solution for point cloud denoising.
• DSD significantly enhances measurement accuracy for annular forgings in noisy environments.
• The method is validated for its effectiveness and practical applicability in industrial applications.