Using Artificial Intelligence to Automate Utilities Asset Inspection & Inventory

Overview

GISonLine, in partnership with the Wrocław University of Technology, are working closely on a project designed to use Artificial Intelligence (AI) and machine learning to process and analyze high resolution imagery to automate utility inspection and asset management workflows.

Integrating AI into the inspection & asset management processes has the potential for significant costs savings and benefits. By using AI and machine learning techniques, utility asset managers and inspectors have the ability to automate the process of identifying and inventorying utility objects and to use the AI algorithms to flag assets recognized as having visible defects or faults.

The Impact

GISonLine is specifically focusing on creating AI algorithms and image pattern recognition for the processing of digital images of railway traction lines obtained from drone and UAV platforms. Traction lines and their connected utility objects are used in support of overhead contact lines associated with electrical power distribution for railway networks.

The formalized process will utilize machine learning techniques to detect and identify the railway traction lines and connected network objects. A primary goal is to train the system for pattern recognition to detect and flag visible defects (faults) found along the traction lines and to automatically alert personnel and inspectors of the issues and maintenance needed.

The Outcome

GISonLine’s objective is to integrate the AI and machine learning technology into its INSPECTonLine cloud-based application.

The INSPECTonLine product will enable asset managers and inspectors with the ability to automate the processes associated with utility asset inventories & inspections and provide a formalized method for detecting and alerting personnel of utility defects that require closer inspection.

Phases of the project include:

1. Analysis of source data and materials (drone & aerial platform data)
2. Definition & characteristics of utility objects to be analyzed
3. Development of traction line recognition algorithms
4. Pattern recognition and feature/fault identification
5. Analysis of the solution for maintaining railway overhead contact lines
6. Integration of the algorithms into the INSPECTonline application

 

INSPECTONLINE   CONTACT US