CLICK HERE if you are having a problem viewing the photos on a mobile device
Using artificial intelligence and their own smarts, two Cupertino teens are working on identifying and predicting areas in a forest that are ripe for wildfires, with the hopes of providing early warnings to fire agencies.
The Smart Wildfire Sensor device by the Monta Vista High School seniors has gotten the attention of Google, and the students have submitted it to Cal Fire for review.
“Our device is unique,” Sanjana Shah said Wednesday. “It combines weather data with real-time classification to predict wildfire.” She and Aditya Shah (no relation, except that they’re schoolmates), both 17 years old, have been working on the device for the past year-and-a-half.
The device uses a camera to take photos in a forest — the students often went to Rancho San Antonio nearby, and sometimes Big Basin Redwood State Park, Sanjana said. Using Google’s TensorFlow machine-learning system, which can process a massive amount of images, the device can then determine the moisture content of the dead fuel — consisting of dead branches and fallen leaves — in a forest. If the moisture content is low, that forest is susceptible to wildfire. The device would then use a wide area network connection to contact fire authorities, the students said.
The pair’s use of the open-source tool TensorFlow landed Aditya a spot to talk about the fire-sensing device at Google’s launch of its AI Impact Challenge — a global contest that’s now taking entries — on October 29.
“One of TensorFlow’s goals is to increase access to AI tools and just seeing what people do with it,” said Brigitte Hoyer Gosselink, head of Product Impact at Google.org, on Wednesday. “Of course, wildfires are very top of mind for all of us in California right now.”
The Camp Fire in Butte County, which started last week, continues to rage and has become the state’s deadliest, claiming dozens of lives so far. It has burned 140,000 acres and was 40 percent contained as of Thursday morning. Another fire, the Woolsey Fire in Ventura and Los Angeles counties, has burned nearly 100,000 acres and is 57 percent contained.
“I’m really thankful to Google for giving me the opportunity to speak about our project,” Aditya said Wednesday. “We were so humbled.”
What Cal Fire does to determine moisture levels now is weigh known fuels, such as a wood stick. That involves human labor, said Jim Crawford, assistant chief of South Bay operations for Cal Fire’s Santa Clara unit, on Wednesday. Aditya and Sanjana’s device would cut down on the need for that human labor by automating the process.
Curtis Brown, head of Cal Fire’s research and development department, said Wednesday that his department is reviewing the teens’ submission.
Sanjana and Aditya, who both count Stanford and UC Berkeley among the colleges to which they would like to apply, have been engaged in using technology to solve other problems, too.
In 2016, Sanjana received the President’s Environmental Youth Award from the Environmental Protection Agency for her creation of a Smart Flood Sensor. She has worked at the Lawrence Berkeley National Lab for the past couple of summers.
Aditya is a coder with a passion for the environment. When he was 15, he started Raindrop US, a nonprofit, to help raise awareness about the California drought. He also started the Green Society at his school to increase awareness about global warming. He has taken summer courses at Harvard, which helped boost his interest in machine learning.
“We have to leverage technology that’s available to us,” Sanjana said. “Machine learning and AI can be shaped today, by our generation, and can be used for social good.”
Might the students be candidates for Google’s AI Impact Challenge, which is taking submissions through January 22? The contest — which includes grant funding from a $25 million pot, plus other Google artificial intelligence resources and expertise — isn’t open to individuals, but Google said it expects to receive entries from plenty of partnerships among students such as Aditya and Sanjana and organizations dedicated to solving different kinds of challenges.
