基于BP神经网络的土壤侵蚀预测模型研究

Research on Soil Erosion Prediction Model Based on BP Neural Network

为提高坡耕地产量,保护水土流失,分析不同下垫面土壤侵蚀量的预测方法,以2000～2001辽宁北部典型坡耕地数据为样本,建立BP神经网络土壤侵蚀量预测模型,应用2002年土壤侵蚀量数据对模型进行检验.结果表明：采用三层BP网络结构,输入层为4个神经元,分别为径流量、降雨量、有机质、覆盖度,输出层为土壤侵蚀量.预测值的合格率为80％,精度较高,具有很好的预测性能.

In order to improve the slope farmland production and protect soil erosion,Analysis of different underlying erosion prediction method is very necessary.In this study,a sample of 2000 to 2001 Liaobei the typical sloping land data,establishing the BP neural network prediction model of soil erosion and data applications in 2002 the amount of soil erosion models were tested.The results show that：The three layers BP network structure,input layer of four neurons,respectively rainfall,runoff,organic matter,coverage,accuracy,output as the amount of soil erosion,simulation is 80％,high precision,has the very good prediction performance.