论文部分内容阅读
糖氮比能够反映作物碳氮代谢的协调程度,及时、准确地监测糖氮比对于作物氮素营养诊断和调控具有重要意义.本研究以不同年份、品种、施氮水平的小麦大田试验为基础,获取鲜叶和粉末状干叶近红外(NIR)光谱及糖氮比信息,分别运用偏最小二乘法(partial least squares,PLS)、BP神经网络(back-propagation neural network,BPNN)和小波神经网络(wavelet neural network,WNN)3种方法建立了小麦叶片糖氮比预测模型,并利用随机选择的样品集对所建模型进行测试和检验.结果表明:小麦鲜叶光谱模型预测性能不佳;而干叶片预测模型表现了较好的准确性,在1655~2378 nm谱区范围内基于3种方法构建的干叶粉末糖氮比估算模型,其预测均方根误差均低于0.3%,决定系数均高于0.9.比较而言,WNN法表现最佳.总体显示,近红外光谱法可以准确预测小麦叶片糖氮比状况,为科学诊断糖氮比提供了理论基础和技术途径.
The ratio of sugar to nitrogen can reflect the coordination degree of crop carbon and nitrogen metabolism, and it is of great significance to timely and accurately monitor the ratio of sugar and nitrogen to the diagnosis and regulation of crop nitrogen nutrition.This study was based on wheat field experiment with different years, varieties and nitrogen levels (NIR) spectrum and sugar-nitrogen ratio information of fresh leaves and powdered dry leaves were obtained. Partial least squares (PLS), back-propagation neural network (BPNN) and wavelet neural (WNN) were used to establish the prediction model of the sugar to nitrogen ratio in wheat leaves, and the model was tested and tested by randomly selected samples.The results showed that the prediction performance of the wheat leaf spectral model was not good, The prediction model of dry leaf showed good accuracy. The root mean square error of prediction of dry matter leaf sugar-nitrogen ratio based on the three methods in the range of 1655 ~ 2378 nm was less than 0.3% The coefficients are higher than 0.9 .WNN method is the best, the overall results show that NIRS can accurately predict the condition of sugar-nitrogen ratio in wheat leaves, which provides a theoretical basis for the scientific diagnosis of sugar-nitrogen ratio Way.