基于“三边”参数的冬小麦冠层SPAD值监测模型

开展冬小麦冠层SPAD值监测,建立“三边”参数与SPAD值之间的高光谱估算模型,以期为高光谱诊断冬小麦冠层SPAD值提供理论依据和技术支持.以冬小麦冠层反射率与冠层SPAD值的相关关系为基础,构建基于“三边”参数的冬小麦冠层SPAD值的一元线性回归模型和主成分回归模型.结果表明:拔节期、抽穗期、灌浆期和全生育期分别以红谷位置、(SD_r-SD_b)/(SD_r+SD_b)、红谷幅值、(SD_r-SD_y)/(SD_r+SD_y)的相关系数最高,且均具有统计学意义(P<0.01);采用主成分方法构建的光谱模型在拔节期、抽穗期、灌浆期和全生育期相较于同期一元线性回归模型,决定系数R²分别提高49.6%,54.3%,14.3%和8.6%,均方根误差RMSE与相对误差RE均分别减少9.0%,12.4%,13.5%和13.6%,因此采用综合光谱信息构建主成分回归模型,在各生育时段及全生育时段对冬小麦冠层SPAD值均有较高的估算精度,可为冬小麦SPAD值的监测与诊断提供依据.

Monitoring model of SPAD value of winter wheat canopy based on″three-edge″parameters

In order to provide theoretical basis and technical support for hyperspectral diagnosis of the SPAD value of winter wheat canopy,the hyperspectral estimation model between″three-edge″parameters and SPAD value was established.Based on the correlation between canopy reflectance and canopy SPAD value of winter wheat,the linear regression model and principal component regression model of canopy SPAD value of winter wheat based on″three-edge″parameters were constructed.The results show that the correlation coefficients of red valley location,(SDr-SDb)/(SDr+SDb),red valley amplitude,(SDr-SDy)/(SDr+SDy)are the highest at jointing stage,heading stage,filling stage and whole growth period,respectively,and all have statistical significance(P<0.01);compared with the same linear regression model,the spectral model constructed by principal component method increase the determinant coefficients R2 by 49.6%,54.3%,14.3%,8.6%at jointing stage,heading stage,filling stage and whole growth stage,and the root mean square error RMSE and relative error RE decrease by 9.0%,12.4%,13.5%and 13.6%,respectively.Therefore,the principal component regression model based on comprehensive spectral information has a high estimation accuracy for the SPAD value of winter wheat canopy in each growth period and the whole growth period,which can provide a basis for monitoring and diagnosing the SPAD value of winter wheat.