不同水分胁迫程度下烤烟叶片钾含量的光谱响应

Spectral response of potassium content in flue-cured tobacco leaves under different degree of water stress

【目的】精准、快速、无损监测烤烟叶片钾素营养水平。【方法】设计不同程度的干旱和渍水胁迫试验,以K326、中烟100为供试品种,探索不同水分胁迫程度下烤烟叶片钾含量与光谱信息的变化规律,筛选烟叶钾含量的敏感光谱特征变量及光谱参数,构建烟叶钾含量预测模型。【结果】(1)随水分胁迫程度的加重,在近红外光区,烤烟叶片钾含量和叶片光谱反射率的值在伸根期均表现为升高趋势,在旺长期与成熟期表现为降低趋势。(2)以利用本文筛选出的8个最佳敏感光谱指数(mSR705、SDr、DVI、MSAVI2、λg、Dr、NDSI(2275,1875)、RDVI)构建的BP神经网络模型效果最好,模型决定系数R2=0.9336,RMSE(均方根误差)为0.1348。【结论】可利用光谱参数构建烤烟钾含量BP神经网络模型,模型稳定、精度较好。可为实时精准监测烤烟叶片钾含量,及时了解土壤水分环境提供技术支撑。

[Objective]This study aims to realize accurate,rapid and nondestructive monitoring of the potassium nutrition level of flue-cured tobacco leaves.[Methods]The experiment of different degree of drought and waterlogging stress was designed.K326 and zhongyan100 were used as the test varieties to explore the change rule of potassium content and spectral information of flue-cured tobacco leaves under different degree of water stress.By screening the sensitive spectral characteristic variables and spectral parameters of potassium content of flue-cured tobacco leaves,the prediction model of potassium content in tobacco leaves was constructed.[Results](1)With the aggravation of water stress,the K content and spectral reflectance of flue-cured tobacco leaves showed an increasing trend at the root extension stage,and a decreasing trend at the vigorous growth and mature stages.(2)The BP neural network model based on the eight best sensitive spectral indexes(msr705,SDR,DVI,msavi2,λg,Dr,NDSI(2275,1875),rdvi)had the best effect,R2(decision coefficient)was 0.9336,RMSE(root mean square error)was 0.1348.[Conclusion]The BP neural network model of potassium content in flue-cured tobacco can be constructed by using spectral parameters and such constructed model is more stable and accurate.This study provides technical support for real-time and accurate monitoring of potassium content in flue-cured tobacco leaves and timely understanding of soil moisture environment.