Two Indian origin High School students develop wildfire prediction AI tool.
Curiosity and Greater Good combination produces the best results, always.
Wildfires have been a menace at many places in the world, specially in California, where they have assumed notorious proportions.
Two Indian-origin high-school students studying at Cupertino, which is home to Apple's headquarters have developed an AI based solution, which can predict the occurrence of wildfires, before take place.
The device called "Smart Wild Fire Sensor" uses Google's open source software Tensor Flow, to combine images and weather parameters to predict the wildfire starts.
Aditya Shah (17) and Sanjana Shah studying at Monsta Vista High School in California; took on this challenge as pat of Google's AI for social good. Aditya has himself experienced several wildfires in his close proximity, since moving here.
Based on the moisture content in the leaves, dryness in the air, , its direction, ambient temperature and density of foliage in the area; the device would predict if the conditions cumulatively thus achieved is conducive to start a fire or not.
The duo have claimed a success rate of over 89 percent, however the challenge remains deploying this solution to wide forests spanning hundreds of thousands of hectares, where potential 'hotspots' could be virtually anywhere.
The other challenge are also to classify a good dataset which could work over a vast forest area, examining dry leaves and twigs of all kinds. Currently Aditya and Sanjana's dataset has only about 500 good quality images.
The students after receiving appreciation and accolades form Google, now plan to take this project to California Fire Department for further support.