中文摘要：

为从区域尺度探讨小麦SPAD的近地高光谱遥感监测技术,采用ASDField Spec 3.0型便携式高光谱仪获取的冬小麦冠层高光谱数据,利用相关分析和偏最小二乘法（PLSR）对SPAD进行建模预测,并采用地统计学方法进行空间变异制图。结果表明,冬小麦叶片SPAD值在不同生长阶段存在一定的差异,但在不同区域之间差异不显著。基于PLSR建立模型,并利用原始光谱和二阶导数光谱进行预测,R~2分别为0.653和0.995,均方根误差分别为2.622和0.327,相对析误差分别为1.549和13.66。综合来看,二阶导数光谱所建立的模型预测能力比原始光谱好。选择拔节期和成熟期进行区域化表达,与实测得到的SPAD空间分布图相比,采用全光谱数据和二阶导数光谱数据预测的SPAD均表现出了较高的空间相似性,其中二阶导数接近实测值。

英文摘要：

Proximal sensing technology has been developed as a useful tool for rapid and economic predictions of many crop physiological and biochemical parameters.Over the past two decades,development of robust and widely applicable visible near-infrared（vis-NIR）diffuse reflectance spectroscopy（DRS）models with an acceptable level of accuracy is an ongoing task.Vegetation SPAD（Soil and Plant Analyzer Development）is one of the most important agronomic parameters（eg.chlorophyll content）for assessing vegetation growth status and health condition.In the present study,the winter wheat canopy hyperspectrum obtained from handheld ASD Field Spec 3.0 were employed to establish the relationships and predication models to assess SPAD content with correlation analysis and partial least squares regression（PLSR）method,and then spatial variability of SPAD in regional scale was characterize by geostatistal analysis method.The result showed that significant differences existed among the four growth stages of elongation,booting,filling and maturation using one-way ANOVA with Tukey-Kramer comparison method at 0.05 level,but there was no significance among the five re-gions（Yuzhou city,Xiangcheng county,Xuchang county,Changge city and Yanling county）.Based on PLSR models,predication by original reflectance R~2= 0.653 was smaller than that by second derivative spectral（R~2=0.995）,with RMSE of 2.622 and 0.327,respectively.Moreover,RPD of the latter was larger（13.66）than that of the former（1.549）.Comprehensive consideration of R~2,RMSE and RPD,model built with the second derivative spectral has better ability to estimate SPAD.Regionalization expression at jointing stage and mature stage showed high spatial similarity for reflectance and the second derivative spectra compared with the measured SPAD.It can be concluded that this method can be extended to determine the physiological and biochemical parameters of winter wheat as well as other crops,with high practical value.