Entel AI and DataRobot use AI to fight forest fires

Friday 8 May 2020

It takes a huge collaborative effort to fight forest fires – and it’s something that is becoming a stark reality for countries like Australia, North America, and Chile. Chile faced an incredibly tough wildfire season recently, after tens of thousands of hectares burned down. But through adversity comes opportunity, especially when it comes to preventing future tragedies from happening.

“Chile is a country that loves its environment. Chilean people live among the trees. The internet of things (IoT) is a powerful medium that can turn just about any object into a data-generating tool. When you attract an internet sensor to a tree, suddenly you have the ability to ‘sniff’ and detect air particles that could potentially help to predict the onset of a forest fire. That’s exactly what EntelOcean did, with the help of DataRobot’s Enterprise AI platform.

“We created an ecosystem of big ideas. We have a nose that can sniff particles. Once you have the particles, you really have to understand what type of particles they are - it could be rain, a barbecue, many things that are actually not a forest fire," says Entel Ocean director Cristobal Urenda Moris.

“We really needed to know when the forest fire was real - that’s when DataRobot came in.” EntelOcean used 'sniffing’ IoT sensors to crunch the data and create an accurate system that could prevent forest files. To do that it had to overcome some major hurdles like eliminating false positives (even something as menial as a raindrop could cause a false positive). Another hurdle involved continuously improving the analytics to create more accurate prediction.

And it worked – according to DataRobot, Entel Ocean is now able to predict forest fires 12 minutes quicker than its previous methods. “Data is huge and it all depends on how you manage every part of it to make it more predictable— seeing things that human beings are not capable of. DataRobot has been a powerful tool to do that," says Ilsun Olmos Ferrufino.

“It was a great moment - we had something here that we could actually use to fight wildfires," adds Cristobal Urenda Moris. Twelve minutes saved is a big deal, considering fire can spread in a matter of seconds. Every minute helps.

“We are focused on using data and AI to get better solutions," says Entel Ocean enterprise digital solutions spokesperson, Lenor Ferrebuz Bastidas. “We have been detecting a forest fire 12 minutes before traditional methods - this is a big deal when it comes to preventing fires.”

Source: itbrief.co.nz

Share |

Copyright 2004-2020 © Innovatek Ltd. All rights reserved.