Big data fusion from hyperspectral and LiDAR sensing for fuzzy-logic-based crop condition assessment in precision agriculture

Large-scale deployment of sensing technologies in precision agriculture generates heterogeneous data streams that are difficult to integrate into actionable crop condition indicators. This study investigates big data fusion from hyperspectral imaging and light detection and ranging measurements for crop condition assessment based on fuzzy logic. Hyperspectral reflectance provides detailed information on biochemical status, while three-dimensional canopy structure derived from light detection and ranging captures geometry and biomass distribution. A scalable fusion pipeline was implemented to aggregate these data into plot-level indicators and to infer crop condition classes through interpretable fuzzy rules. The system was evaluated in multi-crop field trials using independent yield and plant health measurements as reference. Predictive performance of fuzzy-logic-based fusion was compared with hyperspectral-only and deep learning baselines, and spatial patterns of stress detection were analysed at management zone scale. Results demonstrate that combined spectral–structural information significantly improves condition assessment while preserving transparency and expert interpretability.