Minimizing Power Outage For Energy And Utility Industry With Deep Learning

Abstract

To minimize power outages many utility companies, trim the vegetation around their power lines on a fixed time cycle. Weekly updates of satellite images enable detection of trees, changes in tree growth, tree health and tree removal. Using the massive volume of satellite and aerial imagery in combination with deep learning tools, utility companies can gain insight into hazardous trees and tree species around their assets that require spot-wise trimming or removal. Operationalizing satellite-based tree detection and hazard calculation, we present a framework that enable prioritization of vegetation-based hazard removal considering cost, crew size, and the volume of wood that needs to be removed.