Corn yield prediction with artificial neural network trained using airborne remote sensing and topographic data

Artificial neural networks (ANN) are widely used as continuous models to fit nonlinear transfer functions. The objective of the present work was to develop ANN models to predict corn yield from topographic features, vegetation and texture indices. The proposed ANN is back-propagation neural network (BPN) trained by conjugate gradient algorithm. The generalization ability of the best of four models was confirmed by a regression coefficient higher than 90% and a RMSE of 0.365 t/ha, between predicted and observed corn yield.