干旱胁迫下玉米叶片可溶性糖光谱估测研究

Assessing the Soluble Sugar of Maize Leaves in Drought Stress Based on Hyperspectral Data

【目的】利用高光谱遥感技术实时监测干旱胁迫下玉米叶片可溶性糖,为诊断玉米干旱胁迫和田间水分管理提供了理论基础。【方法】采用单因素随机区组设计,对玉米进行盆栽控水,设置4个水分梯度(正常水分、轻度胁迫、中度胁迫和重度胁迫),研究不同干旱胁迫下玉米不同生育时期冠层光谱反射特征及变化规律,比较可溶性糖含量与冠层光谱反射特征和植被指数之间的关系,构建玉米叶片可溶性糖估测模型。【结果】玉米叶片可溶性糖含量在拔节期、喇叭口期和抽雄吐丝期随着干旱胁迫程度的加剧而逐渐增加,均在重度胁迫达到最大值,分别为44.45、44.22和73.00 mg/g,3个时期叶片可溶性糖含量的平均值在抽雄吐丝期达到最大值为72.43 mg/g;不同干旱胁迫下玉米冠层原始光谱反射率在可见光区域(400~700 nm)无明显差异,在近红外光区域(700~1 000 nm)随干旱胁迫加剧逐渐升高,重度胁迫达到最大值;一阶导数光谱的红边幅值随干旱胁迫加深逐渐升高,红边位置在喇叭口期出现"红移";可溶性糖的敏感波段处于原始光谱560~719 nm和导数光谱651~683 nm之间;通过对11个植被指数和经过波段自由组合的RSI、DSI、NDSI这3个植被指数与可溶性糖含量相关分析得知,利用波段自由组合的植被指数DSI(D444,D455)与玉米叶片可溶性糖含量的相关性最好(r=0.99),并由此构建的估测模型y=498 165x2-7 566.9x+71.856能够实现对玉米叶片可溶性糖含量的估测。【结论】不同干旱胁迫下玉米冠层光谱和可溶性糖含量存在差异性,利用高光谱遥感技术可以实时监测玉米叶片可溶性糖含量。

【Objective】The objective of study was to explore real-time monitoring of soluble sugar in maize leaves under drought stress using hyperspectral remote sensing technology provides a theoretical basisfordiagnosisofmaizedroughtstressand field water management.【Method】Single factor randomized block design was used to control the flowering of maize in pots.Four water gradients(normal,mild,moderate and severe)were set up to study the effects of different drought stress on growth,canopy spectral reflectance characteristics and changes of soluble sugar content of corn canopy spectral reflectance characteristics and vegetation index between the construction of maize leaf soluble sugar estimation model.【Result】The soluble sugar content in maize leaves increased gradually with the increase of drought stress during the jointing stage,bell-mouth stage and tasseling stage,and reached the maximum in severe stress,which were 44.45,44.22 and 73.00 mg/g,respectively.The average value of leaf soluble sugar content reached 72.43 mg/g at the time of tasseling.The original spectral reflectance of the canopy did not change significantly in the visible region(400~700 nm),but intensified with the drought stress in the near infrared region(700~1 000 nm).The red edge amplitude of the first derivative spectrum increased with the deepening of drought stress,and the"red shift"occurred at the red edge in the bell mouth stage.The sensitive band of soluble sugar was in the 560~719 nm and the derivative spectra of 651~683 nm.Through the correlation analysis of 11 vegetation indices and RSI,DSI and NDSI,which are free combination of wave band,the three vegetation indices and soluble sugar content,it was found that the vegetation index DSI(D444,D455)had a good correlation with soluble sugar content in maize leaves(r=0.99),and the estimation model y=498 165x^2-7 566.9x+71.856 constructed by this method could achieve the effect on maize leaf soluble estimated content.【Conclusion】There are differences in corn canopy spectra and soluble sugar content under different drought stress conditions.Hyperspectral remote sensing technology can be used to monitor soluble sugar content in maize leaves in real time.