Wildfire Monitoring and Hotspots Detection with Aerial Robots: Measurement Campaign and First Results

Abstract

Wildfires destroy thousands of hectares every summer all over the globe. Nowadays, wildfire prevention is typically done by fire fighters on foot scanning the areas in danger to detect potential hotspots, or by static surveillance systems. These solutions involve either a high human effort, which may also result in dangerous situations, or a high cost to deploy and maintain static surveillance systems. In contrast, aerial robots are a perfect fit for wildfire prevention, as they are able to scan an area autonomously to detect potential hotspots. A review of the available literature has revealed that analysis of hotspot imagery as seen from a drone perspective is absent. In this paper, we describe the first campaign that gathered thermal and visual images with a drone in multiple scenarios. Our findings indicate that even 15 cm hotspots could be easily identified from a drone.