Spring Maize Yield Estimation Based on Combination of Forecasting of Entropy Method and Multi-temporal Remotely Sensed Data

Abstract:

A highly accurate model for crop yield estimation was developed by using the entropy combination forecasting method. Firstly, the single-temporal remotely sensed Landsat TM/ETM+ images at main growth and development stages of spring maize in 2007 and 2008 were used to construct the single-temporal yield estimation models. Secondly, the weights of the single-temporal estimation models were calculated by applying the entropy methods. And then, a combination forecasting model was developed. Finally, the two models were compared. The results showed that the yield estimation model based on combination forecasting and multi-temporal remote images could increase the precision of the yield estimation model based on single-temporal remote images, and the correlation coefficient was remarkably improved in comparison with those of the single-temporal models. They were increased by 0.137 and 0.121 respectively. The values of weights in the combined forecasting showed that the sensitive degree was displayed between main growing stages and maize yield, and that was of great importance for some key aspects: (1) looking for the main limiting factor of maize growth; (2) raising maize yield. Therefore, it is feasible and effective to estimate spring maize yield based on the combined forecasting of entropy method and multi-temporal remotely sensed data. 