Title : Development of an NPK measuring sensor for precision agriculture in Jos, Nigeria
Abstract:
This study investigates the effectiveness of sensor-based technologies in assessing soil nutrient levels, specifically Nitrogen (N), Phosphorus (P), and Potassium (K) compared to conventional laboratory methods within Jos Metropolis, Plateau State, Nigeria. Soil samples were systematically collected from ten georeferenced locations and analyzed using both traditional gravimetric approaches and a portable, IP68-rated soil NPK sensor integrated with an Arduino-compatible RS485 communication interface. The research aimed to evaluate the accuracy, calibration feasibility, and reliability of the sensor device for real-time nutrient monitoring in the context of precision agriculture. Results revealed a consistent trend of overestimation in sensor readings across all nutrient types when compared to laboratory benchmarks: nitrogen (13.10 ppm vs. 1.63 ppm), phosphorus (64.10 ppm vs. 4.24 ppm), and potassium (57.60 ppm vs. 8.02 ppm). Despite this, regression analyses demonstrated strong linear relationships between sensor and lab values (R² = 0.9999 for nitrogen, 0.86 for phosphorus, and 0.78 for potassium), confirming that sensor outputs can be reliably calibrated to reflect true nutrient concentrations. Calibration models were developed using paired sample data, and statistical validation through independent t-tests indicated no significant difference for nitrogen (p = 0.145), though phosphorus (p = 0.014) and potassium (p = 0.033) differences were significant. The study concludes that, while uncalibrated sensors tend to overstate nutrient levels, they offer promising utility for site-specific nutrient management when coupled with localized calibration protocols. These findings contribute to the growing body of knowledge supporting the integration of digital soil sensing into sustainable agricultural systems. Moreover, the work underscores the importance of field-level calibration and hybrid validation frameworks for transforming sensor-based data into actionable agronomic intelligence
