An IoT-Based Soil Nutrient Detection, Monitoring, and Crop Recommendation System for Smart Agriculture in Uganda

Soil nutrient management is essential for improving agricultural productivity and promoting sustainable farming practices. However, many small-scale farmers in Uganda have limited access to affordable and timely soil testing services. This paper presents an Internet of Things (IoT)-based soil nutrient detection, monitoring, and crop recommendation system designed to support data-driven farming decisions. The system integrates an NPK sensor with a Raspberry Pi 3 Model B using an RS-485 communication module to collect real-time Nitrogen, Phosphorus, and Potassium values. The collected data is processed using Python, stored in a MySQL database, and displayed through a web-based dashboard developed using HTML, CSS, PHP, and Chart.js. A Random Forest machine learning model is deployed through a Flask API to recommend suitable crops based on detected soil nutrient levels. Experimental evaluation conducted on a 100 m² agricultural plot in Soroti, Uganda showed reliable data transmission, acceptable sensor accuracy, and strong usability. The system achieved 99.8% data transmission success, sensor mean absolute error between 3.2% and 4.1%, and 90% machine learning model accuracy. The proposed system provides an affordable precision agriculture solution for smallholder farmers.