Inspiration
Wildfire distress in both Los Angeles and Australia.
What it does
Creates a predictive map that shows how a wildfire will spread over a geographic region for the next 20 minutes based on wind speed.
How we built it
We built a home-made wind gauge, and used computer vision to map how fast its spinning. Using this data we inserted relative wind speeds into a predictive model that uses the Rothermal equation (predicts how wildfires spread).
Challenges we ran into
getting the math to be correct and make sense, accurate computer vision for our wind gauge, and making an appealing User Interface for the program.
Accomplishments that we're proud of
A well-polished interface that is simple and very effective.
What we learned
How to make computer vision more accurate for faster moving objects, and how to integrate multiple variables into a predictive model.
What's next for WindMesh
Making a physical mesh of nodes that can implement wind gauge data into the simulation automatically and more accurately.
Log in or sign up for Devpost to join the conversation.