利用高光谱遥感技术, 分析晚播条件下小麦叶片与冠层模式光谱特征和叶面积指数（LAI）的变化规律, 建立了适用于晚播小麦的叶面积指数估算方法。 研究结果表明: （1）从红光和蓝紫光420～663 nm波段提取的叶绿素光谱反射率植被指数（CSRVI）与旗叶SPAD值做相关性分析, 结果表明正常播期和晚播处理在叶片模式的相关系数分别为0.963*和0.997**, 达显著和极显著水平。 （2）利用相关性分析, 得出两个播期处理的LAI与SPAD值相关系数分别是0.847*和0.813*, 均达到显著水平。 SPAD值与LAI及CSRVI指数均具有相关性, 可以用CSRVI指数建立LAI的估算模型。 （3）对叶片模式和冠层模式光谱曲线特征分析得出, 叶片模式中在680～780 nm处的反射率呈现陡升趋势, 在可见光波段的446和680 nm和近红外波段的1 440和1 925 nm处各有两个明显的吸收波谷, 在540～600, 1 660和2 210 nm波段处有两个明显的反射波峰; 三种冠层模式中60°模式下的光谱反射率整体表现为最高。 （4）将各波段反射率与叶面积指数做相关性分析得出在可见光波段范围内, 光谱反射率与LAI总体呈现负相关性, 500～600 nm处有一个波峰。 （5）将三种冠层模式下（仪器入射角度分别与地面呈30°, 60°和90°夹角）的等效植被指数与LAI做相关性分析得出: 60°冠层模式下八种植被指数与正常播期LAI的相关性均未达显著水平, 比值植被指数（RVI）、 归一化植被指数（NDVI）、 增强型植被指数（EVI）、 再次归一化植被指数（RDVI）、 土壤调整植被指数（SAVI）、 修改型土壤调整植被指数（MSAVI）的等六种植被指数与晚播条件下的LAI具有显著和极显著相关关系; 90°冠层模式下CSRVI指数与正常播期处理的LAI具有显著相关关系, NDVI指数与晚播处理的LAI具有显著相关关系; 30°冠层模式下的八种植被指数与两播期处理的LAI的相关性均未达显著水平。 综合分析CSRVI指数、 NDVI指数的相关性最高, 这两种指数最具有估算LAI的潜力。 （6）通过三种冠层模式所计算的植被指数估算LAI模型, 结果表明, 正常播期条件下, 其最佳估算模型是90°冠层模式CSRVI指数所建立的线性模型Y=-7.873 6+6.223 8X; 晚播条件下的最佳模型是60°冠层模式RDVI指数所建立的幂函数模型Y=30 221 333.33X17.679 1, 两个模型的决定系数R2分别为0.950*和0.974**。 研究表明试验中所提取的CSRVI指数能够反映旗叶叶绿素含量, 可以通过光谱仪器的叶片模式对小麦生育期内叶绿素含量进行监测; 通过冠层模式计算的CSRVI指数和RDVI指数所建立的LAI估算模型可以对小麦的LAI进行无损害观察。

In this study, hyperspectral remote sensing technology was used to measure the changes of leaf and canopy characteristics with leaf area index (LAI) of wheat leaves under late sowing conditions, and the LAI estimation method suitable for late sowing wheat was established. The results show that: (1) Correlation analysis between chlorophyll spectral reflectance vegetation index (CSRVI) is extracted from the red and blue bands (420～663 nm) to analyze the correlation between SPAD value and CSRVI of the leaf mode under normal sowing and late sowing treatment with R2 being 0.963* and 0.997** reached significant and extremely significant level, respectively. (2) It is concluded that the correlation coefficients of LAI and SPAD values for the two sowing dates are 0.847* and 0.813* by using correlation analysis, respectively, and both reaching significant levels. The SPAD value is correlated with LAI and CSRVI indices, and the CSRVI index can be used to establish the LAI estimation model. (3) Analysis of the spectral curves of characteristics of leaf pattern and canopy patter shows that the reflectance of leaf pattern increases sharply at 680～780 nm. There are two distinct absorption troughs at 446 nm, 680 nm in visible light band and 1 440 and 1 925 nm in near-infrared wave band. There is a clear reflection peak at 540-600 nm band. There are two distinct reflective peaks at 1 660 and 2 210 nm, and the spectral reflectance of the three canopy modes is the highest in the three canopy modes. (4) Correlation analysis between the reflectance of each band and the leaf area index shows that the spectral reflectance has a negative correlation with the overall LAI in the visible light range, and there is a peak at 500～600 nm. (5) Correlation analysis of the equivalent vegetation index and LAI in the three canopy modes (the angle of incidence of the instrument with the ground at 30°, 60°, and 90° respectively) is obtained: there was no significant correlation between 8 vegetation indices and the LAI under the late sowing condition of 60° canopy mode. And a significant and extremely significant the 6 vegetation indices (normalized vegetation index (NDVI), enhanced vegetation index (EVI), re-normalized vegetation index (RDVI), Soil-adjusted vegetation index (SAVI) and modified Soil-adjusted vegetation index (MSAVI) ) under the late sowing condition of 60° canopy mode; the CSRVI indices in the 90° canopy mode were significantly correlated with the LAI of the normal sowing date. NDVI index is significantly correlated with LAI in late sowing treatment; the correlation between the 8 vegetation indices in the 30° canopy mode and the LAI in the two sowing dates was not relevant. Comprehensive analysis of the CSRVI index, NDVI index is the most relevant, and these two indices have the most potential to estimate LAI. (6) The LAI model was estimated by the vegetation index calculated by the three canopy models. The results show that under the normal sowing date, the best estimation model is the Linear function model established by the 90° canopy model with CSRVI index Y=Y=-7.873 6+6.223 8X; The best model under late sowing conditions is the power function model Y=30 221 333.33X17.679 1 established by the 60° canopy mode RDVI index, with R2 being 0.950* and 0.974** in the two treatments, respectively. Studies have shown that the CSRVI index extracted from the test can reflect the chlorophyll content of flag leaf. The chlorophyll content of wheat during the growth period can be monitored by the leaf pattern of the spectroscopy instrument; LAI estimation model based on CSRVI index and RDVI index calculated by canopy model can be used to observe wheat LAI without damage.