Forest fire prevention and control agencies in São Carlos, a city in the interior of São Paulo, Brazil, will soon benefit from advanced aerial technology to detect and combat fires more rapidly.

Researchers at the São Carlos Engineering School of the University of São Paulo (EESC-USP) are developing drones equipped with gas sensors and artificial intelligence to identify forest fires in their early stages. The project was introduced during the aeronautics session of FAPESP Week Toulouse, held from June 10 to 12 in Toulouse, southern France.

The team is collaborating with the Civil Defense, São Carlos city government, and the local Department of the Environment. A proposal has been submitted to evaluate the drones for identifying fire hotspots in the municipality. According to Glauco Augusto de Paula Caurin, professor at EESC-USP and project coordinator, the drones are outfitted with compact, cost-effective sensors that continuously measure carbon dioxide, methane, temperature, and humidity in the air.

These sensors function collectively like an electronic nose, analyzing gas concentration data using AI systems. The technology enables the detection of emission sources by identifying gases typically released during forest fires. This airborne monitoring provides much faster detection than satellites, allowing authorities to act swiftly before fires grow out of control.

Over recent years, the team has evaluated the use of drones for monitoring greenhouse gas emissions. The results have shown that drones are both efficient and more affordable compared to current methods like satellites, research aircraft, or observation towers. Unlike satellites, which may only pass over an area every two days, drones can be deployed multiple times and at specific locations of interest.

Another benefit of using drones is their ability to collect data at varying altitudes. This allows researchers to generate three-dimensional measurements of gas concentrations, offering volumetric insights rather than just surface-level data. This kind of information is not currently obtainable even with the most advanced satellites.

Despite these advantages, current commercial drones are limited to flight times of 15 to 30 minutes, making them less suitable for large-scale forest coverage. To address this, the research group is working on aerodynamics projects to improve drone efficiency, enabling longer flights and broader area coverage.

Initial tests were conducted around the USP campus in São Carlos, a region situated between the Atlantic Rainforest and the Cerrado biome. The researchers aim to expand their missions to the Amazon in the future, once the drone technology becomes more advanced and capable.

By Impact Lab