Drones and AI can help suppress wildfires in the future

Drones and AI can help suppress wildfires in the future

In the FireMan research project, drones, machine vision and AI are trained to help rescue professionals suppress fires and reduce the damage they cause. Agile aerial drones can detect the signs of a fire from far away and transmit data to rescue departments in real time.

Members of the research project are the Finnish Geospatial Research Institute (FGI) of the National Land Survey of Finland (NLS), VTT Technical Research Centre of Finland Ltd, and the Universities of Jyväskylä and Oulu.

Methods tested using real fires

Data on restoration burning by Metsähallitus has been collected in the project to teach AI. Drones have been used to video forests before, during and after burning, and AI has been trained to reliably detect smoke and the advancement of a fire.

In September, the technology was tested in practice in the Emergency Services Academy Finland's training area in Kuopio - however, forest floor mats delivered to the area were burned instead of actual forests.

During the tests, a drone's AI detected a fire from a video feed and sent an alarm. The project employed Finland's first drone fleet developed for wildfires, which independently located and detected the fire and monitored it from different angles. The data collected by the drones was transferred in real time to a situation centre established in the project, where an accurate digital model was used to predict the spread of the fire and plan firefighting measures. An independently moving drone was also used in the tests, which is able to navigate in a dense forest without being operated by a person.

The drones teamed in Kuopio with Probot Oy's ground robot, which can transport equipment or participate in rescue operations in difficult environments, including forests, in which regular equipment is ineffective. The use of transport robots may be part of the firefighting measures of the future.

'We tested the effectiveness of the drones and AI-based systems developed in the project successfully in realistic conditions,' says Eija Honkavaara, Professor at FGI who leads the project.

'We were happy to see that these methods really worked. This strengthened our understanding that drones and AI can be key in the early detection and suppression of wildfires. We also identified challenges, for example, with the network connections of drones. These can be addressed in future research projects,' she says.

While technologies help professionals, they cannot replace people

Airspace management is developed in the project to keep drones separate from other airspace users, including planes and helicopters. In Finland, VTT is developing a U-space concept to enable the safe operation of drones in the airspace with other aircraft. The U-space system was demonstrated during the tests in Kuopio. In the future, similar solutions will enable the use of drones in crisis areas, for example, so that manned and unmanned aerial vehicles can operate in the same airspace without any risk of a collision.

'Technologies will not replace rescue professionals, but they will make rescue operations quicker and easier. For example, data transmitted by drones can help transport firefighting equipment and personnel more rapidly to where they are most urgently needed,' says Hannu Karvonen, Senior Scientist at VTT.

The FireMan project is funded under the EU's Recovery and Resilience Facility, which promotes the green and digital transitions through research. The funding is coordinated by the Academy of Finland.

Lisätietoja 

Eija Honkavaara
Research Professor
Finnish Geospatial Research Institute of the National Land Survey of Finland
[email protected]
029 531 4716

Hannu Karvonen
Senior Scientist
VTT
[email protected]
040 021 6396

Ilkka Pölönen
Associate Professor
University of Jyväskylä
[email protected]
040 024 8140

Tuomo Hänninen
Research Director
University of Oulu
[email protected]
0294482845

• UNITE Flagship (uniteflagship.fi)
• 6G flagship: 6G Flagship - More than wireless