冬小麦生物量敏感波段的提取对冬小麦长势指标监测具有极其重要的意义.通过测定冬小麦生物量和冠层光谱,采用相关性分析,偏最小二乘法和逐步多元线性回归法进行建模与分析,最终提取了冬小麦生物量敏感波段.结果表明,利用偏最小二乘法实现了数据的降维,利用多元统计分析提取的冬小麦生物量特征波段为470,570,870,895,1 170,1 285,1 355,1 360 nm,所有敏感波段集中在可见光和近红外波段,其中,近红外波段最多,并基于提取的敏感波段构建了冬小麦生物量的监测模型,模型的验证参数TC为0.229,R2达到0.809,RMSE为0.351,取得较高的估算精度.为了验证模型的稳健性和准确度,利用野外大田数据进行验证,结果表明,实测值与预测值具有较高的拟合性,其拟合系数达到0.87.研究结果表明,利用多元分析方法对冬小麦敏感波段提取是可行的,可为冬小麦生物量监测提供一定的理论和实践参考.%Extracting the sensitive bands of winter wheat biomass is of importance in monitoring the growth indicators of winter wheat.In the study,the biomass of winter wheat and its canopy spectral reflectance were measured,and their relationship was analyzed using the multivariate statistical analysis.Finally,the model of winter wheat biomass was constructed and validated based on the sensitive bands extracted with the method of multivariate analysis.The results showed that dimension of hypersectral data was reduced,and the sensitive bands of winter wheat biomass extracted with multivariate analysis were 470,570,870,895,1 170,1 285,1 355,1 360 nm.All the sensitive bands covered in the visible and near infrared region where most of sensitive bands concentrated in the near infrared bands.Moreover,the monitoring model for winter wheat biomass was built based on the sensitive band extraction and the result demonstrated that the validation parameters of monitor model were 0.809,0.229,and 0.351 for R2,TC,and RMSE,respectively,which indicated that the monitor model achieved a higher estimation precision.To further verify the robustness and accuracy of the biomass model and the sensitive bands,the monitor model was validated using another experiment which the eco-climate differed with other two experiments,and the fitting degree between the measured and the predicted value reached to 0.87.The study showed that it was feasible to extract the sensitive band of winter wheat using multivariate analysis method,and it would provide some theoretical and practical references for monitoring the biomass of winter wheat in future.
展开▼
