Digital Asset Capturing (DAC): an image recognition based algorithm supporting collection and analysis of overhead power grid assets

Abstract

Since last decade, electrical power companies have largely improved quality and efficiency of overhead power lines surveillance and infrastructure assets inspections thanks to aerial vehicles (AVs) applications. Regular visual inspection are performed to check the status of their power lines mainly using helicopter equipped with external gimbals housing cameras to detect anomalies or abnormal conditions as soon as possible. Current detailed aerial inspection approach uses trained inspectors who fly aboard helicopters to inspect the lines with binoculars and cameras, while recording the data in a logbook. These manual activities are expensive and affected by human errors. The paper presents a different approach to digital capture power grid assets and to detect visible anomalies on components based on Artificial Intelligence technologies.