We wanted to focus on the biggest phenomena or systems that had the biggest impact on the environment. After some research, we learned the importance of honeybees in detail and we also learned how murder hornets represents a clear and present danger to the North American honey bee population. We decided to do something about it.
What it does
We created a system that is able to identify the presence of murder hornets at a beehive and shoot them down with a 5W diode laser. This information is then relayed to a website that can communicate with beekeepers, government officials, and researchers which beehive locations are presently being attacked or may be in danger within the next 24 hrs.
How we built it
Using an Arduino with two server motors and a powerful camera, the turret is able to visualize the target and aim the laser on the murder hornet. The machine learning algorithm behind the camera is able to identify which flying objects are the targets. This information is sent to a flask-powered website built with html/css.
Challenges we ran into
Training the model to identify between the hornets vs. bees required several attempts to reach a satisfactory level of precision. Getting the laser to shoot in the precise location of the hornet was another challenge which required patience and practice. Connecting the front and backend was yet another challenge, as members of the team wanted to expand their skills into either front or backend that they had never worked with before.
Accomplishments that we're proud of
Members being able to switch positions (i.e. backend vs. frontend) in less than 24 hrs was something we've very proud of (and glad it worked out!). We were all out of our comfort zones at some point today! Creating a more complete project, with hardware, backend, and frontend is also an accomplishment none of us had achieved as of yet.
What we learned
We learned how to be flexible with the software we chose to work with, as well as how to train an algorithm with a dataset. We also learned how to create frontend graphics and content using flask and connecting that with the backend.
What's next for Hornets Bee Gone
Improving the quality of the aiming of the laser is the biggest goal. This would be done by using a microphone to differentiate the frequency of wing beating between the murder hornets and the honeybees. This, along with reinforcement learning and some supervised learning algorithms, would be crucial to improving accuracy and increasing safety.