DETECTION AND SUPPRESSION OF AN INCIPIENT FIRE USING A UAV

Abstract

This thesis investigates the development of an Uncrewed Aerial Vehicle (UAV) system designed to detect, localize, and suppress incipient wildfires. The system is designed with a commercial UAV in mind, but with an emphasis on the scalability and portability of its components. The key contributions of this research are the design and testing of a multi-spectral fire detection algorithm utilized in broad area surveys, a fire localization routine that utilizes the fire detection data to design an optimized flight path for revisiting potential fire sites, and the integration of a third-party drop system with the commercial UAV in order to deliver suppressant payloads to fire sites. These critical components are combined together to form an end-to-end fire process chain, which fuses autonomous and manual UAV functions into a procedure where one UAV can conduct fire detection, localization, and suppression activities utilizing two different mission profiles. Extensive testing of the fire process chain in totality and across its individual components is conducted, the results are thoroughly analyzed, and suggestions for future wildfire management UAV systems are provided.