Anyone who sees an unusual flying object near Möhrendorf in Middle Franconia these days is probably witnessing the Evolonic pilot project. When the weather is good, a drone equipped with a camera scans forest areas fully automatically. Their mission: to detect the source of a fire at an early stage and thus prevent it from getting worse. Researchers from the Friedrich-Alexander University of Erlangen-Nuremberg and the Fraunhofer Institute for Integrated Systems and Device Technology (IISB) are involved in the project.
“If the drone detects that smoke is rising from the forest, it immediately sends a live image to the fire control center responsible, including position determination and information about the terrain, for example whether it is mountainous or flat and whether there are buildings or technical systems that are particularly worthy of protection in the vicinity are,” says Tobias Raczok from the IISB. The control centers could then quickly order a suitable team and suitable equipment to the site. The earlier a fire source is detected, the easier it is to extinguish the fire and prevent it from spreading.
It is true that the fires in Hawaii in particular are making dramatic headlines, but due to climate change, wildfires are to be expected more and more frequently, even in moderate latitudes. Last year, fires in Germany destroyed more than 3,000 hectares of forest – an area about the size of the island of Borkum – and with it valuable habitat for plants and animals. The stored carbon is also released as carbon dioxide, further fueling climate change.
If the plan of the researchers from southern Germany works out, autonomous drones could help prevent such large-scale infernos of flames in the future, and also quite cheaply. Raczok reports that sports pilots, who used to control the forests with their planes during hot periods, are more expensive and not always available everywhere.
Distinguish between smoke and dust
The current test drone is three by three meters in size, weighs ten kilograms, has four propellers and additional wings like a glider. The propellers propel it vertically upwards until it switches to gliding flight at an altitude of 30 to 40 meters. Your target height is 120 meters. In order to detect the smoke, it is equipped with an optical camera and uses AI-supported software for the evaluation.
During a joint exercise with the Erlangen fire brigade, the team collected further data for the drone’s image recognition algorithm.
(Image: Tobias Raczok)
“We trained the AI with forest fire recordings that fire brigades from all over Germany made available to us,” explains the researcher. The recordings came from helicopters and squadrons of drones, as well as from private drones and small aircraft. According to Raczok, the biggest challenge is distinguishing between smoke and dust that is thrown up during field work. This is one of the reasons why a thermal imaging camera is to provide additional data in the future and secure the evaluation of the optical images.
Over the next two years, the drone will scan 80 square kilometers of forest for four months at a time. It takes about an hour and a half to fly over this area. As a rule, they get by with one battery charge, says Raczok. Tests with smoke bombs and real fires should show how well the smoke detection actually works. However, the approval for this is still pending.
Image 1 of 7
A team of researchers from Friedrich-Alexander-Universität Erlangen-Nuremberg and the Fraunhofer Institute for Integrated Systems and Device Technology (IISB) developed the autonomous drone.
(Bild: Sven-Nicolas Evens – Evolonic)
The issue of security is particularly important to the Evolonic team. “Nobody should be harmed. That’s why we work with redundant systems,” emphasizes the Fraunhofer researcher. The drone can not only communicate via mobile communications, but also via satellites and via a direct radio link. In addition, it sends and receives signals, such as those intended to prevent collisions in manned aviation. The researchers want to further optimize systems that can recognize paragliders, hot-air balloons and gliders. According to the researcher, birds are not endangered. “The drone is much larger than a bird and is therefore avoided.”
In the future, the team wants to work even more closely with biologists and the forest authorities. After all, if the drone is trained accordingly, it can not only detect fires, but also bark beetles, for example – and thus provide further important information from the forest.
To home page
#Prevent #fires #Drone #automatically #scans #forest #areas