Automated feeder routing for underground electricity distribution networks based on aerial images

Justus Ameling, Gunther Gust

Abstract: Highlights•Deep neural network encodes geoinformation from aerial images.•Integrated A* search identifies effective feeder routes based on this encoding.•Evaluation based on real-world study using more than 7000 actual feeder segments.•Feeder segments generated are on-par or superior than competing approaches.•Approach facilitates operational and strategic network planning.