Estimation of Winter Wheat Plant Nitrogen Concentration from UAV Hyperspectral Remote Sensing Combined with Machine Learning Methods

Nitrogen is one of the most important macronutrients and plays an essential role in growth development winter wheat. It very crucial to diagnose nitrogen status timely accurately for applying a precision management (PNM) strategy guidance fertilizer field. The main purpose this study was use three different prediction methods evaluate wheat plant concentration (PNC) at booting, heading, flowering, filling, whole stage Guanzhong area from unmanned aerial vehicle (UAV) hyperspectral imagery. These include (1) parametric regression method; (2) linear nonparametric (stepwise multiple (SMLR) partial least squares (PLSR)); (3) machine learning (random forest (RFR), support vector (SVMR), extreme (ELMR)). also pay attention impact stages on accuracy model. results showed that compared with regression, method could evidently improve estimation PNC, especially using SVMR RFR, training set model flowering filling explained 93% 92% PNC variability respectively. testing 88% 91% variability, root mean square error validation (RMSEtesting) 0.82 1.23, relative deviation (RPD) 2.58 2.40, Therefore, conclusion drawn it best choice estimate UAV Using methods, respectively, achieve outstanding performance, which provide theoretical basis putting forward PNM strategy.