膜下滴灌棉花根系发育参数的BP模型预测

Prediction of root growing parameter of drip-irrigated cotton under mulch by using BP model

通过对人工神经网络理论中BP网络的分析,建立描述不同水分条件下膜下滴灌根冠间非线形变化的模拟模型,通过该模型利用棉花地上部株高、干物重、叶面积参数、时间及土壤水分预测膜下滴灌不同水分处理下的根系参数,以2000年在新疆的大田棉花试验结果作为学习样本和检验样本。结果表明,所建立的人工神经网络模型对描述不同水分条件下根、冠复杂的非线性关系方面具有较高的精度和应用价值。

By analysis of BP network of artificial neural network, a simulation model to describe the nonlinear relations between root and canopy of crops different water states was established. The model is successful in predicting root parameters of drip-irrigated cotton in different water states based on the cotton height beyond the land, dry weight, leaf area index, time, and soil water. The results of field research at Xinjiang in 2000 indicate that the model of artificial neural network is of high practical value and high precise for the simulation of complicated nonlinear relations between root and canopy.