Cocotech: Iot and AI-Driven Solutions for Coconut Farm Automation

Abstract

The coconut industry in Sri Lanka faces challenges due to rainfall inconsistency and soil conditions. This research aims to develop machine learning (ML) algorithms to predict coconut yields and market prices using historical data from Coconut Research Institute (CRI) of Sri Lanka and Central Bank (CB). The models are trained, tested, and assessed on a monthly dataset, with their performance estimated using Mean Squared Error (MSE) and Mean Absolute Error (MAE). The results suggest that sophisticated ML methods can improve decisionmaking in the coconut industry, leading to better production planning and market strategies. The research also presents an IoT and ML-based system using ESP32 sensors for live monitoring and Random Forest for drying prediction. A decision-supporting mobile app ensures consistent copra quality, improved efficiency, and maximum oil yield. The system also explores intelligent water scheduling for coconut cultivation, providing precise irrigation recommendations. Also, this research includes Copra fungus detection and copra grading using ML technology and image processing as one of the system functionalities