Dale Faubion

The VineTech project helps vintners accurately predict grape yield using images collected throughout the growing season and processed through a machine learning pipeline.

My Contributions

I modified the FarmNG Amiga Rover by adapting legacy hardware/software to function on the new platform.

I solved a major system limitation where the rover reset to factory defaults on every power cycle. By implementing a non-standard workaround, I preserved configurations and enabled persistent system updates.

System Integration

Using OpenGoPro, I enabled Bluetooth communication between the rover’s Ubuntu-based brain and multiple GoPro cameras. I also helped develop an onboard application allowing users to start and stop image capture in the field.

Machine Learning

The system uses a model developed by Dr. J. Walker Orr to analyze vineyard images segmented by bays and rows, achieving an approximate 10% error rate in yield prediction.