摘要：Cover crop is an agriculture operation that is planted during the winter and owns several advantages such as improving water quality and soil quality. However, the large-scale effect of cover crop in relieving environmental burden and improving cash crop yield over a region has not been widely investigated. Due to cost and time limitation, it is not favorable to conduct the conventional field trials. Previous study proposed a Random Forest classifier to predict the pattern of cover crop adoption from the remote sensing data. In this study, we propose a Multilayer Perceptron neural network to further improve the performance and reliability of the classification model and achieve an accuracy of 0.93 and Cohen’s Kappa of 0.76. Moreover, the Multilayer Perceptron model outperforms two baseline classification models. Finally, we predict the cover crop planting status for the Knox County and found a significant increase in cover crop planting on the corn cropland in 2016.
