Skip to Content

I2L: Intelligent Power Line Inspection

Union Fenosa Distribucion, Prysma, INTA and Diagnostiqa, funded under INPACTO Program Spanish Ministry of Economy and Competitivity IPT-2012-0491-120000




The I2L project has the goal of developing new technologies to optimise high-voltage power line inspection and maintenance. The general aim of the project is to automatise as much as possible the whole process of acquisition and processing of the aerial inspection data in order to achieve a rapid and efficient power line inspection.

The Computer Vision Group is in charge of the Work Package on Automatic Data Processing. The main responsibility of the CVG group is to conduct research and develop novel strategies and algorithms for processing data from visual, infrared and LIDAR sensors, on-board aerial vehicles, for automated fault detection in and around powerline infrastructure. The research efforts of CVG are focused on automating the power line inspection process by looking for strategies that satisfy the different requirements of the inspection: detection of transmission towers in videos, insultator detection and identification of the faulty/defective one and the analysis of the security distances. 

The solutions developed by CVG fall under three broad modules: 

  • Module I: LIDAR data analysis and processing for powerline inspection;
  • Module II: Automatic tower detection and tracking in the visual and infra-red data;
  • Module III: Vision-based electrical insulator fault detection


Module I: LIDAR data analysis  


The key advantage of the LIDAR data over the visual information is the availability of the depth information. A LIDAR sensing equipment samples the surface (e.g. forests, pipelines, powerlines and other infrastructure) and produces a highly accurate 3D representation of the surface (a point cloud). For powerline inspection, such points provide sufficient information for detecting some very specific type of faults/anomalies, such as faults in structures, sagging cables and violations of the powerline corridor.

Key Objective: To identify all entities (violations) in the powerline corridor that are causing or can potentially cause damage to power cabels and towers.

A solution based on LIDAR data processing: 

LIDAR data Analysis LIDAR processing pipeline

Algorithms for automatic classification of the LIDAR data into:

  • power cables;
  • transmission towers;
  • powerline corridodr; and
  • violations

software toolbox for:

  • pointcloud visualization and manipulation; and
  • interface to algortihms for pointcloud classification


Demonstration of the software toolbox and LIDAR pointcloud processing and analysis algorithms:


Module II: Automatic tower detection and tracking
Module III: Vision-based electrical insulator fault detection

 A fault detection system has been developed for the automatizing the processing of data issued from intensive visual inspection of insulators. This system consists of an insulator localization algorithm detecting the individual plates composing the insulator and a fault detection algorithm processing each of the detected plates.

These algorithms have been tested on data issued from intensive aerial inspection in various regions of Spain and showing various environmental conditions such as forests, plains, mountainous and urban surroundings. The algorithms have proven robust to many disturbances such as complex surroundings, important insulator point of view variation, two distinct insulator materials (ceramic and tempered glass), various sources of damage (among which bird excrements, rust, flashover consequences and chipped plates) and partial occlusion. The system is capable of localizing 92% of the insulator plates successfully, and predicts correctly the state of a plate (OK or faulty) in 99.6% of the cases. 

Demonstration of insulator detection under partial occlusion:





Starting date: 
August 2012
Finishing date: 
June 2016