Partition dynamic threshold monitoring technology of wildfires near overhead transmission lines by satellite

Wildfires are a major natural disaster that can threaten the safe and stable operation of overhead transmission lines. Compared with large-area forest fires, transmission-line wildfires usually cover a small area and spread rapidly, making monitoring accuracy and real-time requirements of high priority. Wildfire monitoring based on satellite remote sensing has advantages in terms of monitoring-range width and the capacity for real-time monitoring; however, the detection threshold changes dynamically due to the influences of climate, geography, weather, and other factors that affect monitoring accuracy. To focus on small-area wildfires near overhead transmission lines, we developed a partition dynamic threshold calculation method based on time-series prediction. Basic thresholds are obtained based on a large number of historical values, followed by partitioning one of these values according to digital elevation model data and subsequent correction. Compared with conventional constant-threshold monitoring methods, our proposed method significantly reduced missed and false detection rates. Additionally, to improve fire-spot localization to the overhead transmission-line towers, we developed a tower-location algorithm based on block searching. Compared with the traditional traversal algorithm, our algorithm enabled a 15,000-fold increase in operation speed. These improvements will significantly enhance the monitoring of transmission-line wildfires, which are highly reliant upon alarm speed.