Using AI-driven Drones to Put Out a World on Fire
In their research paper, “Towards early forest fire detection and prevention using AI-powered drones and the IoT,” eight scientists from universities in Abu Dhabi, Malaysia, and Turkey sought to find ways to use two of the latest technologies – artificial intelligence (AI) and drones (or uncrewed aerial vehicles (UAVs) – to find new ways to detect wildfires fires sooner and predict ways to best manage the risks and more quickly extinguish them.
Using a unique UAV-based IoT (UIoT) approach, the team developed a low-cost, low-maintenance fire-detecting tool to map a given terrain and facilitate large-scale firefighting . On the technical side, the units use a long-range wide area network (LoRaWAN) outfitted on an autonomous drone equipped with a camera. A cloud back-end server facilitates monitoring, analysis, and reporting. The result: fires can be detected in 1-5 minutes after a fire ignites.
The new system was able to classify fires with 99.46% accuracy and to a precision rate that, on average was 99.64%. Consequently, a novel method has been devised to reliably recognize wildfires such that battery power consumption is minimized, thus increasing battery lifespan.
Where there’s smoke, now there’s an AI drone-enabled fire detection system
Once a fire has been detected, an UAV outfitted with a camera, flies to the location where the fire is located. A tracking algorithm then rapidly devises the best countermeasures to distinguish the fire. Static gateways connect neighboring sensing nodes and integrate them.
This paper proposes a UAV-IoT system design for wildfire detection. Our system is based on low-cost and low-maintenance fire-detecting IoT nodes that communicate using LoRaWAN and HTTPS communication protocols with our cloud server to report the sensed measurements. Once a fire is detected at a specific location, an autonomous drone with a visual camera flies toward the identified geolocation of the node to monitor the fire’s progress. The drone runs our novel fire center detection and tracking algorithm for precise localization and then uses an appropriate countermeasure to extinguish the fire. The proposed system uses static gateways connecting neighboring sensing nodes architecture for system integration. Authorities receive detailed real-time reports on the fire’s location, progress, and the countermeasures taken. Numerical results show that the UAV-IoT system can be a cost-efficient alternative to satellite imaging for wildfire detection, especially when the cost of fire-relevant losses is high.
Keep up with the latest AAM news, insight and analysis.
Top 3 Takeaways
- AI-equipped drones can drive down the cost of fighting wildfires
- AI-equipped drones with cameras can more quickly and accurately predict where firefighting resources will be needed and thus facilitate better fire management
- Combining the IoT and UAVs is a winning formula for wildfire detection and mitigation
- #internet of things (IoT) #LoRaWAN network #wildfire, #AI, #fire detection #uncrewed aerial vehicles (UAV) #autonomous navigation
#drones #aam #aam today #firefighting #AI #artificial intelligence #uav #uavs #evtol #evtols
Follow Us
