基于动态模糊神经网络模型预测水稻受砷胁迫下叶绿素的含量变化

Prediction of Chlorophyll Content in Rice under Arsenic Stress Based on Dynamic Fuzzy Neural Network Model

摘要: 为探讨利用动态模糊神经网络模型评价水稻受重金属砷污染胁迫状况，研究了水稻农田在自然生长环境下重金属砷污染对水稻叶片中叶绿素含量的影响，并做出对叶绿素含量变化敏感的植被指数与叶绿素之间的相关性分析，通过多元逐步回归找到与叶绿素含量变化敏感的植被指数NDVI、MNDVI、MTCI、MSR、GNDVI作为动态模糊神经网络的输入参数，水稻叶片中叶绿素含量值作为模型的输出参数建立能够判断水稻污染等级的动态模糊神经网络模型。结果表明，预测的叶绿素含量值与实测的叶绿素含量值拟合度高(R² = 0.905)。说明可以用动态模糊神经网络模型预测农田的污染等级情况，为大面积监测水稻农田污染提供借鉴的依据。

Abstract: The aim of this paper is to assess the ability of dynamic fuzzy neural network model in evaluating the pollution status of rice field under the stress of heavy metal arsenic pollution. Heavy mental arsenic pollution would change the chlorophyll content in leaves of rice, so the correlation between chlorophyll content and vegetation index was analyzed. By stepwise regression, NDVI, MNDVI, MTCI, MSR, GNDVI were found which were very sensitive to the change of chlorophyll content. Then a dynamic fuzzy neural network was constructed where the five indices were used as input parameters and the chlorophyll content of rice as the output parameter. The results reflected that the predicted chlorophyll content is in good agreement with the measured chlorophyll content (R² = 0.905). It was indicated that the chlorophyll content of rice under the stress of heavy metal arsenic pollution could be predicted by dynamic fuzzy neural network model, which would provide a reference for large area monitoring of heavy metal pollution in rice field.