Enhanced Feature Extraction From Assimilated VTCI and LAI With a Particle Filter for Wheat Yield Estimation Using Cross-Wavelet Transform

To further reveal the relationships between different variables and yield at each growth stage of winter wheat, an approach for estimating regional yields wheat multiple time scales was developed by assimilating CERES-Wheat model simulations remotely sensed observations. Specifically, particle filter assimilation algorithm chosen to assimilate simulated soil moisture depth 0–20 cm leaf area index (LAI) MODIS retrieved vegetation temperature condition (VTCI) LAI. The resonance periods series assimilated VTCIs LAIs stages with crop were analyzed separately using cross-wavelet transform determine variation scales, calculated weights VTCI LAI used establish a estimation model. Both could comprehensively integrate effects sensing observations, transformed had specific yields, regardless whether they or not. Compared unassimilated variables, given greater key stages, namely jointing heading-filling milk maturity enhancing feature extraction accuracy improved.