Total nitrogen estimation in agricultural soils via aerial multispectral imaging and LIBS

Abstract Measuring soil health indicators (SHIs), particularly total nitrogen (TN), is an important and challenging task that affects farmers’ decisions on timing, placement, quantity of fertilizers applied in the farms. Most existing methods to measure SHIs are in-lab wet chemistry or spectroscopy-based methods, which require significant human input effort, time-consuming, costly, low-throughput nature. To address this challenge, we develop artificial intelligence (AI)-driven near real-time unmanned aerial vehicle (UAV)-based multispectral sensing solution (UMS) estimate TN agricultural farm. macro-nutrient SHI directly crop health. Accurate prediction can significantly increase yield through informed decision making timing seed planting, fertilizer timing. The ground-truth data required train AI approaches generated via laser-induced breakdown spectroscopy (LIBS), be readily used characterize samples, providing rapid chemical analysis samples their constituents (e.g., nitrogen, potassium, phosphorus, calcium). Although LIBS was previously for nutrient detection, there no study integration with UAV imaging AI. We two machine learning (ML) models including multi-layer perceptron regression support vector predict using a suite classes characteristics crops red (R), near-infrared, green (G) spectral bands, computed vegetation indices (NDVI), environmental variables air temperature relative humidity (RH). generate training models, determine N spectrum (collected from farm) calibration model correlation between actual maximum intensity spectrum. In addition, extract features images captured while follows autonomous flight plan, at different growth stages crops. ML model’s performance tested fixed configuration space hyper-parameters various hyper-parameter optimization techniques three wavelengths