2025 · RAIL Phase II Robotics Cohort II Challenge, KNUST
Asaase TomGuard
A robot that detects tomato diseases in the field.
My role: Instructor and technical supervisor
What it is
Detect tomato diseases early, in the field, and alert the grower.
What I did
- Trained and mentored the Opoku Ware M/A JHS robotics team, who competed with this build.
- A four-wheel-drive robot with a camera, a Raspberry Pi and an Arduino, and a Streamlit dashboard that raises alerts.
- Fine-tuned two YOLOv8n detectors, one for tomato leaves (9 classes) and one for tomato fruit (8 classes).
- In October 2026 rebuilt the robot software: a checksummed UART link between the Pi and the Arduino, a row patrol state machine and the Arduino firmware, tested against a Python twin of the firmware. Not yet run on the robot.
Outcome
- 1st place, JHS category, RAIL Phase II Robotics Cohort II Challenge (January 2026), won by the Opoku Ware M/A JHS team.
- Validation mAP50 of 0.931 on leaves and 0.760 on fruit.
Built with
- Python
- YOLOv8
- Streamlit
- Raspberry Pi
- Arduino
Recognition
- Jan 2026: 1st Place, JHS category, RAIL Phase II Robotics Challenge, RAIL KNUST