Vineyard Yield Prediction Machine Learning Project
The VineTech project is designed to help vintners (vineyard owners) to accurately predict how many tons of grapes
will be harvested from the vineyard via images collected throughout the growing season which are sent through a
Machine Learning algorithm.
Utilizing the FarmNG's Amiga Rover, our team first reduced the track width to a record low of just under 30 inches,
then got to work on adapting the legacy system's software and hardware to the new rover.
A major challenge I solved was overcoming a safety measure built in by FarmNG Robotics, which reset the files and
configurations of the rover to a factory default each time the rover is power cycled, thus wiping any updates and
changes to the rover we performed. After finding a non-standard workaround, we were able to get the right BlueTooth
services running to allow the rover to connect to the GoPros used for image collection. Our team developed an app
on the Amiga's Brain (built in Ubuntu computer) which allows any user to start and stop capturing images in the field.
The project utilizes a Machine Learning algorithm developed by Dr. J. Walker Orr, one of our project advisors. The
algorithm receives images divided into bays and rows, outputting a prediction of how many pounds of grapes will be
produced by each bay of vines, which has proved an error rate of only 10%.
Projects and what I'm up to these days:
Cybersecurity Competition Score Cards
- 550th (top 6%) National Cyber League Individual Spring Game (2025) - Score Report (pdf)
- 51st (top 1%) National Cyber League Team Spring Game (2025) - Score Report (pdf)
- 23rd (top 11%) National Cyber League CyberSEED Team Game (2025) - Score Report (pdf)
- 1350th (top 16%) National Cyber League Individual Fall Game (2024) - Score Report (pdf)
- 152nd (top 3%) National Cyber League Team Fall Game (2024) - Score Report (pdf)