Fire responders have limited resources to fight large scale forest fires: they are using inefficient methods and tools to plan and allocate those scarce resources. They forecast fire behavior and spreading using old methods and cumbersome tools that are like driving by looking in the rearviewmirror.
This can be grealy improved by leveraging UAVs, machine learning and small-factor computers.
Forest fires are sadly here to stay. Most recently, bushfires in Australia and the Amazon forest fires gained notoriety, but other areas like Central Africa and Patagonia have large forest fires every year.
These fires are destroying an extremely valuable natural resource: forests make up for around 30% of the Earth's surface, and they mitigate the impact of climate change as well as support the livelihoods of 250 million people.
The 40% of the extreme rural poor
75% of global biodiversity
Forests sustain more than three quarters of the global biodiversity. The uncontrolled spreading of forest fires creates unsurmountable damage to the wildlife and a severe unbalance of the ecosystem.
We are developing tools to fight this
Data collection with UAVs, data analysis with machine learning and affordable small computers need to be incorporated into well-designed, simple-to-use tools for fire response coordinators.
Fire boundary detection
Our machine learning model detects the coordinates of fire spots and recognizes the fire boundaries, giving responders crucial information to coordinate the firefighting efforts.
Forecasting wildfire behavior
Multispectral imagery, machine learning algorithms and a portable weather station combined, enable us to build an accurate, local and almost real-time forecast of wildfire spreading.