Advancing Precision Agriculture Through Integration of Unmanned Aerial Vehicle Imagery and Ground Sensor Data in Northern Nigerian Farms

Precision agriculture in Nigerian smallholder systems remains nascent, limited by high technology costs and absence of locally calibrated methodologies. This study advances precision agriculture through integrating UAV-derived multispectral imagery with in-situ soil sensor data across six research farms in Kano, Kaduna, and Niger states. A workflow for generating high-resolution within-field variability maps of soil nutrients, crop stress indicators, and weed infestation pressure using UAV RGB and multispectral imagery is developed and evaluated across four cropping seasons. Machine learning regression models relating UAV-derived spectral indices and textural features to soil-measured variables achieve mean prediction errors of 12 percent for soil organic carbon and 18 percent for available phosphorus. A variable rate input application algorithm translates within-field variability maps into spatially differentiated fertiliser and pesticide recommendations. Economic analysis of variable rate application versus uniform management demonstrates an average net benefit of 14 percent across the study farms. The study develops a simplified precision agriculture protocol adaptable to resource constraints of Nigerian extension services and farmer cooperative organisations. Keywords: precision agriculture, UAV, smallholder, soil variability, Nigeria