Google’s AI Research: Part 2

Digital representation of wildfire growth, overlaid on Google Maps

In the previous post we did a deep dive on Google’s FloodHub; their proprietary software used for modeling floods and sending alerts to affected communities. Google has also been hard at work on detection and alerts for another kind of natural disaster: Wildfires.

Smoke from wildfires seen from space

You can see that detecting these fires from space is quite a bit easier than getting accurate information on flooding using satellite imagery. That is unless the wind shifts and holds the smoke directly over top of the fire in an inversion layer.

Put simply Google can put together a few polygons of where the fire is located and label them red for fire based on changes in the image, or green for normal. Then they can build a fire boundary based around these polygons and figure out what communities nearby may be at risk. Thermal imagery can also be used to map out wildfire boundaries.

Fighting wildfires is risky business – you have highly trained firefighters working in the bush for weeks at a time, and the situation can quickly change such that the environment becomes too dangerous for them to continue working. Once a fire starts and reaches a critical size, we can only do so much. Often times are best option is containment, limiting the damage, and praying for better weather. A lot of work needs to be done on the prevention and risk management side since we don’t really have a good handle on controlling them. This happens at all levels of government – from federal aid in the form of the military and wildfire management, to local governments adopting regenerative agriculture. Wildfire suppression runs the gamut from satellite imagery in space to beavers (yes, the animal).

Part of the future of helping with climate change should include a focus on where we build our cities. Throughout history we’ve seen entire civilizations completely wiped out by changing climate, so building a lasting city that never has to move is actually not as normal as we think. Greater care should be taken to avoid areas prone to natural disasters.