基于BP神经网络与GIS可视化的作物需水量预测

Prediction of Crop Water Requirement Based on BP Neural Network and GIS Visualization

作物需水量是农田水利工程规划、设计与灌溉用水管理的重要参数。基于BP神经网络与GIS可视化作物需水量预测,以河南省冬小麦的作物需水量为例,利用实测站点的地理坐标以及高程,通过BP神经网络对分析数据进行加密插值,并采用GIS里的协克里金法结合高程作为一个协变量进行作物需水量的空间插值,实现GIS可视化,从而得到准确度较高的作物需水量空间分布图。结果表明,该方法对于河南省冬小麦作物需水量预测具有较好的预测精度,方法有一定的参考价值。

Crop water requirement is an important parameter of farmland water conservancy project planning,design and irrigation water management.Prediction of crop water requirement based on BP neural network and GIS visualization to predict crop water requirement.By using geographic coordinates and elevation,encryption interpolation of winter wheat crop water requirement in Henan Province,through BP neural network,and spatial interpolation by the method of cokriging combined with the association elevation,higher accuracy spatial distribution is obtained.The results show that the method for winter wheat in Henan Province predicting crop water requirement of the space has a great degree of accuracy in prediction,the method has some reference value.