基于指数平滑算法组合模型的玉米实时变量追肥研究

Research on Real-time Variable Topdressing of Corn Based on the Combination Model of Exponential Smoothing Algorithm

玉米中耕时期是追施氮肥适宜阶段,为了进一步提高肥料利用率,简化变量施肥作业流程,实现变量施肥采集到施肥过程整体性,提高中耕变量追肥实时性,提出了一种动态多模型指数平滑算法组合。算法实现NDVI数据实时计算处理与趋势预测,根据历史数据实时修正减少计算误差,实时动态调整算法的平滑系数,从而间接缩短变量施肥数据处理响应时间,修正误差以提高作业精度。仿真分析该算法结果表明:在数据波动变化情况下,数值平滑程度以及趋势均优于单一指数平滑法,并且实时得到下阶段NDVI的准确预测趋势发展,计算结果精度高,为变量施肥提供可靠的理论施肥量,及时根据作物需求补充氮肥减少差空间异性。

Cultivation period was a suitable stage for topdressing nitrogen fertilizer.To simplify the workflow from collection to fertilization and increase fertilizer utilization and improve the real-time performance of variable topdressing in cultivating,a dynamic multi-model exponential smoothing method combination was proposed to realize real-time processing calculation and trend prediction of NDVI data.The algorithm corrected the calculation error in real time based on historical data,and dynamically adjusted the smoothing coefficient of the algorithm in real time.The algorithm indirectly shortened the response time of variable fertilization data processing and corrected the collection error.Simulation analysis of the algorithm showed that in the case of data fluctuations,the degree of numerical smoothing and trend were better than the single exponential smoothing method.The system obtained the accurate prediction value of NDVI in the next stage in real time,and the calculation result of the algorithm was highly accurate,which provided a reliable theoretical fertilization amount for variable fertilization.In the process of variable fertilization of corn,nitrogen fertilizer should be added in time according to the needs of crops.