Computer Engineering Thesis
VERDE.
Intelligent hydroponic monitoring and decision-support system integrating environmental sensors, camera-based monitoring, and an application for lettuce cultivation.
Conceptual architecture only. Exact implementation connections should be confirmed against the project repository and paper before publication.
SensingpH · EC/TDS · turbidity · water temperature · humidity monitoring
→
Raspberry PiAcquisition, filtering, validation, and logging
→
Monitoring appSystem status, readings, and monitoring results
Camera inputPrepared lettuce image datasets
→
YOLOv8nHealthy · yellowing · wilting predictions
→
Decision supportInterpretation for cultivation monitoring
What the team built
Data with context.
The system combines sensor acquisition, filtering, validation, and logging with a computer vision component and a monitoring application.
The team prepared datasets, trained a YOLOv8n model, and evaluated preliminary predictions for healthy, yellowing, and wilting lettuce.
My contribution
Integration work.
Hands-on work included hardware connections, sensor interfacing, calibration, data acquisition, filtering, validation, logging, and system troubleshooting.
Preliminary examples
Confidence is not accuracy.
Example prediction confidences shown in the supplied brief: Healthy 90% / 99%, Yellowing 93% / 95%, Wilting 71% / 73%.
These are individual preliminary example predictions—not overall accuracy, precision, recall, or mAP. The Random Forest water-quality classifier was not yet loaded in the supplied application screenshot.
Prototype photos were not present in the VERDE folder.
Use the thesis PDF below for the supplied technical documentation.
Use the thesis PDF below for the supplied technical documentation.