HotSpot is a machine learning tool that uses computer vision to detect wildfires in aerial images captured by planes, drones, satellites, etc. This was inspired by California's wildfire crisis in August 2020.
Wildfires are often noticed by humans after they have gotten too large and too close to civilization. By detecting these wildfires from the air before they get too large and reach humans, HotSpot enables authorities to take action early to contain and put out the wildfires, prevent and mitigate damage, and save countless lives.
Check it out here: https://hotspot-asathiya.herokuapp.com/. Enjoy!
The repo containing the code for the the PyTorch neural network used for detecting wildfires from aerial images (from drones, planes, satellites, etc.) and the Flask API that exposes this network can be found here: https://github.com/asathiya007/hotspot-api.
The repo containing the React user interface code can be found here: https://github.com/asathiya007/hotspot.