Highlights
This was a multidisciplinary build where the main challenge was not one algorithm or one board. It was making several technical layers behave as one useful field system.
The problem statement involved geo-tagging plantation growth and spotting early crop stress using aerial imaging. Our solution used a drone carrying RGB and NIR capture, GPS-tagged imaging, and machine-learning assisted detection on a Jetson Nano pipeline.
My contributions were hands-on and system-level. I assembled the drone, tuned PID behavior in Betaflight for stable flight, worked on firmware-side integration, and helped connect the ML pipeline with the imaging and GPS data flow.
Winning the competition mattered, but the more important thing was learning how to contribute effectively inside a cross-domain team where software, control, sensing, and deployment all intersect.