Abstract: Nitrogen content is one of the important nutritional indicators of wheat. The traditional method of monitoring nitrogen content of wheat leaves in different growth stages was complicated and destructive. Therefore， a method for rapid prediction of nitrogen content in wheat using remote sensing images of unmanned aerial vehicle （UAV） was proposed in this study. The UAV images acquired during the wheat filling period were used for splicing and fusion based on the improved weighted average algorithm，which could effectively eliminate splicing marks and differences. Thus，wheat features were effective extracted，including RGB features，HIS features and vegetation indices（VIs），which were used to construct support vector regression（SVR）model. The experimental results showed that the accuracy of the HIS-based+VIs combination model was the highest. The validation set coefficient （R²） was 0. 774 and the root mean square error（RMSE） was 0. 363 7. The results indicated that the monitoring method based on UAV for the nutrition of wheat during grain filling was feasible， which also provided technical support for the management of wheat field.

Key words: filling stage of wheat, unmanned aerial vehicle（UAV）, support vector regression, nitrogen content