为准确估算作物需水量,提高水分利用效率,采用RBF神经网络预测参考作物腾发量,由于参考作物蒸发蒸腾量的影响因子很多,且各影响因子间的相关性很大,运用主成分分析的原理,将影响参考作物蒸发蒸腾量的因子降低维数。以山西省某灌区的参考作物腾发量为例,运用DPS软件找出了3个综合因子来代表众多因子并作为RBF人工神经网络的输入,运用Matlab7.0进行编程,对参考作物腾发量进行预测。结果发现其预测结果与用Pen-man-Monteith公式算得的值具有很高的一致性,与BP神经网络相比,RBF神经网络具有学习速度快等优点,将此方法用于参考作物腾发量的预测可以收到理想的效果。
In order to accurately estimate crop water requirement,improve water use efficiency,the RBF neural network is used to predict reference crop evapotranspiration in this paper.Because the reference crops evapotranspiration is affected by many high correlated factors,the principal components analysis(PCA) is used to reduce the dimension of the reference crops evapotranspiration affecting factors.In this paper,the reference crop evapotranspiration of an irrigation district in Shanxi Province is taken for example,the DPS software is used to identify three factors to represent the number of integrated factors,which are used as the input of the RBF artificial neural network,and the matlab7.0 programming is used to predict the reference crop evapotranspiration.The result shows that its forecasting result has very high uniformity with that of Penman formula.Comparing with the BP neural network,the study speed of RBF neural network is more quickly.Using this method to predict the reference crop evapotranspiration can obtain ideal effect.