AI is being leveraged in countless ways, from medicine to self-driving cars. Now, researchers at the University of South Australia are leveraging it to help detect and prevent forest fires. The team’s research is published in a recent article in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.
The dangers of forest fires have increased in recent years, largely due to a changing climate that has led to dry, hotter seasons. You’ve probably seen numerous news stories about blazes ravage millions of acres of land uncontrollably. Indeed, as a result of a warming climate, researchers have seen a nearly two-fold increase in forest fire activity between 1984 and 2015. With a constantly warming climate, wildfire activity is only expected to get worse, posing significant to humans and animal life alike.
At the University of South Australia, researchers are turning to AI technology for potential solutions to detect these wildfires before they can happen, which may give responders time to slow their spread before they can become fiery infernos.
Specifically, researchers have leveraged cube satellites that are designed to detect fires from space. While there has historically been a challenge with obtaining and processing the vast amount of hyperspectral information these satellites can detect and transmit, the addition of AI technology may help as a game changer. In fact, researchers estimate that AI-equipped satellites may be able to detect wildfires quicker than with data gathered on the ground.
Specifically, these satellites use gathered hyperspectral imagery (eg, light reflected from the Earth) to create hyper detailed images of the planet’s surface, which is designed to help with monitoring for wildfires and even other land management tasks.
Leveraging AI tools, researchers found that they could reduce how much energy was consumed to download and analyzed data, and was able to detect smoke up to 500 times faster than on ground tools. In this case, AI can help detect fires just from smoke, which means it can detect growing wildfires before a blaze even starts.
Sources: Science Daily; NOAA; IEEE
