Power Line Extraction and Tree Risk Detection Based on Airborne LiDAR

Area of Science:

• Electrical Engineering
• Computer Vision
• Forestry Science

Background:

• Safe operation of transmission lines is crucial for human activities.
• Timely identification of tree risks is essential for preventing power outages and ensuring grid stability.

Purpose of the Study:

• To propose an automated method for power line extraction and tree risk detection.
• To enhance the safety and reliability of transmission line operations through accurate risk assessment.

Main Methods:

• Utilized height difference and local dimension feature probability models for power line point extraction.
• Employed Cloth Simulation Filter algorithm and neighborhood sharing for conductor and ground wire classification.
• Applied linear-catenary model for conductor reconstruction and safety distance calculation for tree risk grading.

Main Results:

• Achieved high classification accuracy for conductors and ground wires, with precision, recall, and F-score exceeding 98%.
• Demonstrated accurate conductor reconstruction with low Root-Mean-Square Error (<3.67 cm) and Maximum Error (<7.13 cm).
• Validated the effectiveness of the tree risk detection method with a Mean Absolute Error of <6.47 cm in safety distance calculation.

Conclusions:

• The proposed method effectively extracts power lines and detects tree risks with high accuracy and reliability.
• This approach contributes to improved transmission line safety and operational efficiency.
• The findings support the integration of advanced computational methods in power infrastructure management.