Inspiration
Motivated by the advent of the amazon wildfires in recent times coupled with an article on the use of drones in planting trees across the US. We wanted to investigate the use of AI for social good, but on a global scale and investigated techniques to improve efficiency of planting saplings via use of AI for drone vision.
What it does
Unity communicates via a Python bridge to native C++ code and Computer Vision techniques evaluate mean free path estimation on the images from virtual drone cameras to plant sapling across a simulated terrain.
How we built it
Hopes, dreams and tears. Unity and AirSim were the main frameworks - without which this wouldn't have been possible.
Challenges we ran into
Started at 7pm Saturday, whoops. Drone wouldn't move. Oops.
Accomplishments that we're proud of
The drone flies and plants seeds, this is awesome!
What we learned
Use of AirSim and not to use experimental Libs.
What's next for Ride of the Amazonian Valkyries
Extension to swarm behaviours. Improvement of AI to use imitation learning strategies to use reward functions. Correct for environmental factors.
Log in or sign up for Devpost to join the conversation.