Vegetation management is the work of finding and trimming the trees that threaten a utility's power lines, before one falls and causes a power cut or a fire. Satellite images and AI now help utilities decide which trees to deal with first.
The problem is that a utility may own lakhs of kilometres of lines. Sending crews along every line to look at the trees is slow and expensive. Most of the trimming is done on a fixed cycle, whether or not a tree is a danger. So some trees are cut that did not need it, and a fast-growing one can touch a line between two visits.
The software uses pictures instead. Earth observation satellites photograph the land along each line. Geospatial AI reads the pictures, finds the trees, measures their height and how close each one is to a line, and notices which are growing fastest. It marks the lines on a map from a geographic information system (GIS). The result is a ranked list of the spans where trees are most dangerous. A span is the stretch of line between two poles or towers.
This software is bought mainly by utilities in the United States, where a tree touching a live line can start a wildfire. So the example here is American. Consider a utility in northern California with lines through dry forest. Before the summer fire season, the software shows 40 spans where trees are within 3 metres (about 10 feet) of a line, and marks 6 of them as urgent because the trees are tall and leaning. The utility sends its trimming crews to those 6 first. It also sends the list to the control room, so operators know which lines are most likely to trip in a storm.
The pictures themselves are analysed by satellite imagery analytics, which turn raw images into answers. In India, the usual harm from a tree on a line is a power cut in the monsoon or in a storm. Indian utilities mostly still trim on a fixed cycle, but the same satellite method can be used for their lines too.