Topdressing Nitrogen Demand Prediction in Rice Crop Using Machine Learning Systems

This research is an outcome of the R&D activities Ecodevelopment S.A. (steadily supported by Hellenic Agricultural Organization—Demeter) towards offering precision farming services to rice growers. Within this framework, a new methodology for topdressing nitrogen prediction was developed based on machine learning. Nitrogen key element in culture and its rational management can increase productivity, reduce costs, prevent environmental impacts. A multi-source, multi-temporal, multi-scale dataset collected, including optical radar imagery, soil data, yield maps monitoring 110 ha pilot farm Thessaloniki Plain, Greece, four consecutive years. RapidEye imagery underwent image segmentation delineate zones (ancillary, visual interpretation unmanned aerial system scenes employed, too); Sentinel-1 (SAR) modelled with Computer Vision detect inundated fields (through this) indicate exact growth stage crop; Sentinel-2 data were used map leaf concentration (LNC) exactly before applications. Several learning algorithms configured predict various levels, XGBoost model resulting highest accuracy. Finally, curves select dose maximizing yield, which thus recommended grower. Inundation mapping proved be critical process. Currently, expanding application method different study areas, view further empower generality operationality.