Towards the Automation of Wildfire Monitoring with Aerial Vehicles: The FIREFRONT Project

This paper describes an integrated system for wildfire monitoring with aerial vehicles. The system is composed of an airborne payload compatible with UAVs and manned aircraft and associated communication and software infrastructure. The system is able to detect fire and smoke in the payload thermal and RGB images, georeference their location in ground coordinates using telemetry (GPS, IMU), and forecast the fire front spread. We describe the progress made in the payload system, acquired datasets, fire/smoke segmentation methods, fire georeferencing, and wildfire spread modeling, providing the reader with a overview of the project and related technical publications.