MINI-UAV LIDAR FOR POWER LINE INSPECTION

Geer Teng, Mei Zhou, C. R. Li, Haohao Wu, Wenqiang Li, Fanjiang Meng, Chengcheng Zhou, Lian Ma2017 · DOI

TL;DR

This paper introduces AOEagle, a compact UAV-borne LiDAR system designed for efficient and safe power line inspection, particularly in challenging mountainous terrain. The system's effectiveness is demonstrated through real-world data, showcasing its ability to extract critical information like power line corridors, detect pole deformation, and identify potential hazards.

What Problem Does This Paper Address?

Traditional power line inspection relies heavily on manual surveys, which are time-consuming, costly, and pose safety risks, especially in difficult-to-access areas like mountains. While existing solutions like ground-based LiDAR and airborne laser scanning (ALS) have limitations in terms of weight, cost, or accessibility, this paper addresses the need for a flexible, portable, and cost-effective solution. Specifically, it targets the gap in efficient and accurate power line inspection in complex terrains where traditional methods fall short.

Methods

The core of this paper is the introduction and application of the AOEagle system, a mini-UAV-borne LiDAR developed by the Academy of Opto-Electronics, Chinese Academy of Sciences. This system is mounted on a multi-rotor UAV. The methodology involves:

  • Data Acquisition: Conducting test flights in mountainous terrain with the AOEagle system.
  • Point Cloud Processing: Utilizing commercial software (Inertial Explorer 8.60 and TerraScan) for processing the raw LiDAR data. This includes:
  • PPP Processing: Employing precise point positioning for accurate georeferencing.
  • Calibration: Ensuring accurate sensor calibration.
  • Point Cloud Computing: Generating dense 3D point clouds.
  • Feature Extraction and Analysis: Applying algorithms to extract specific features from the point clouds for inspection purposes, including:
  • Corridor and line point cloud extraction.
  • Pole and transmission line point cloud extraction.
  • Pole deformation detection (leaning, bending).
  • Span and sag detection for transmission lines.
  • Dangerous point detection (identifying potential hazards like trees, buildings).

Datasets

The paper utilizes real-world data collected during a test flight in Xichang, Sichuan, China, on March 14, 2017. This dataset consists of LiDAR point clouds from a power line inspection mission. No new benchmarks are introduced; rather, the paper validates the system's performance against existing standards and field measurements.

Key Insights and Ideas

  • Mini-UAVs as a Game Changer: The paper strongly advocates for mini-UAVs as an ideal platform for LiDAR, offering a balance of portability, flexibility, and cost-effectiveness, especially for complex terrains.
  • AOEagle's Versatility: The AOEagle system proves capable of extracting a wide range of critical information for power line inspection, from basic geometric measurements to detailed structural analysis.
  • Accuracy in Challenging Environments: Despite operating in mountainous terrain with varying altitudes, the system achieves high accuracy in measurements like pole height and span distance, comparable to traditional surveying methods.
  • Safety and Efficiency Gains: The automated extraction of features and identification of potential hazards significantly enhances the safety and efficiency of power line inspection compared to manual methods.

Key Results

  • Corridor and Line Extraction: The system successfully extracts power line corridors and the transmission lines themselves, enabling preliminary inspection of operational surroundings and hidden dangers.
  • Pole Deformation Detection: Pole leaning and deformation can be detected with high accuracy. For example, the measured pole height (53.3317m) was within ±0.02m of the total station measurement (53.35m), satisfying the angle-steel pole leaning resolution requirements (0.5%).
  • Span Detection: The span detection accuracy was confirmed with a relative error of 1/7773 compared to field observations.
  • Dangerous Point Detection: The system can identify potential safety hazards by measuring minimum distances between transmission lines and ground features.

Limitations and Future Work

  • Limitations:
    • The maximum ranging distance of 200m can lead to some ground points being lost, especially in complex terrain.
  • The paper notes that the distance between the camera and the GPS antenna needs to be at least 30cm to avoid signal interference.
  • Future Work:
    • The authors suggest exploring the integration of AOEagle with other devices, such as CCD cameras, for enhanced data acquisition.
  • They also propose investigating AOEagle's performance in other application areas, including 3D city modeling and forestry resource management.
  • Further research into integrating AOEagle onto other platforms like cars is also mentioned.

Summary updated

Abstract

Light detection and ranging (LIDAR) system based on unmanned aerial vehicles (UAVs) recently are in rapid advancement, meanwhile portable and flexible mini-UAV-borne laser scanners have been a hot research field, especially for the complex terrain survey in the mountains and other areas. This study proposes a power line inspection system solution based on mini-UAV-borne LIDAR system–AOEagle, developed by Academy of Opto-Electronics, Chinese Academy of Sciences, which mounted on a Multi-rotor unmanned aerial vehicle for complex terrain survey according to real test. Furthermore, the point cloud data was explored to validate its applicability for power line inspection, in terms of corridor and line laser point clouds; deformation detection of power towers, etc. The feasibility and advantages of AOEagle have been demonstrated by the promising results based on the real-measured data in the field of power line inspection.
BibTeX
@inproceedings{teng2017mini,
  author = {Geer Teng and Mei Zhou and C. R. Li and Haohao Wu and Wenqiang Li and Fanjiang Meng and Chengcheng Zhou and Lian Ma},
  title = {{MINI-UAV LIDAR FOR POWER LINE INSPECTION}},
  booktitle = {˜The},
  year = {2017},
  doi = {10.5194/isprs-archives-xlii-2-w7-297-2017},
}
Figure