Step feature line extraction from large-scale point clouds of open-pit mine based on structural characteristics

Area of Science:

• Geosciences
• Computer Science
• Mining Engineering

Background:

• High-precision step feature lines are vital for various open-pit mine operations.
• Existing methods struggle with the complexity and detail of step feature lines in large-scale point clouds.

Purpose of the Study:

• To develop a novel technique for accurate step feature line extraction from open-pit mine point clouds.
• To improve the efficiency and precision of mining planning and slope maintenance.

Main Methods:

• K-dimensional tree (KD-tree) resampling and bilateral filtering for point cloud preprocessing.
• Point Cloud Properties Network (PCPNET) for normal estimation, slope, and aspect calculation.
• Morphological operations for refining slope lines and creating a continuous dataset.

Main Results:

• The proposed SFLE_OPM method achieved an accuracy score of 89.31% and a true positive rate of 80.18% on the OPMSFL dataset.
• Demonstrated superior extraction accuracy and precision compared to existing methods.
• Introduced the Open-Pit Mine Step Feature Line (OPMSFL) dataset for benchmarking.

Conclusions:

• The SFLE_OPM technique effectively extracts complex step feature lines from large-scale open-pit mine point clouds.
• The developed method offers a significant advancement in mining engineering applications.
• The new dataset and method contribute to more precise mine design and management.