基于光谱指数的古尔班通古特优势种沙蒿叶绿素含量的估算模型

A model for Estimating Chlorophyll Content of Sand Sagebrush of Gurbantongout Dominant Species Based on Spectral Indexes

叶绿素是监测植被生长健康状况的重要指标之一,高光谱遥感技术可以为植物叶绿素含量的定量化监测提供一种简便有效且非破坏性的采集方法。利用ASD光谱仪测定沙蒿的反射光谱曲线,计算国内外常用的13种光谱指数,分析光谱指数与沙蒿叶绿素的关系,选择并建立能指示沙蒿叶绿素含量的指标和最优模型。结果表明:简单比值指数(SR)、色素简化指数(PSSR)、归一化植被指数(NDVI1)和(NDVI2)、结构不敏感色素指数(SIPI)、归一化比值叶绿素指数(PSND)等6个高光谱植被指数在生长过程中均与沙蒿叶绿素含量有较好的相关关系,相关系数在99%置信水平下均达到0. 9以上;在6种植被指数建立下的估算模型中,决定系数大小依次为RSIPI>RPSND> RPSSR> RNDVI2> RNDVI1> RSR,决定系数均达到0. 9以上,其中光谱指数SIPI的决定系数最高为0. 983,说明根据SIPI所建立的模型能够很好地发挥预测性,适用于沙蒿叶绿素的估算。

Chlorophyll is one of the important indexes for monitoring vegetation growth and health.The hyperspectral remote sensing technology can provide a simple, effective and non-destructive method for quantitative monitoring of plant chlorophyll content. This study used the ASD spectrometer to determine the reflectance spectra of sand sagebrush, and calculated the 13 spectral indices commonly used at home and abroad.It analyzed the relationship between spectral indices and chlorophyll of sand sagebrush, while it chose and established an index and an optimal model to indicate the chlorophyll content of sand sagebrush. The results showed that in the growth process, six hyperspectral vegetation indices, including Simple Ratio Index （SR）, Pigment Simplification Index （PSSR）, Normalized Vegetation Index （NDVI1,NDVI2）, Structural Insensitive Pigment Index （SIPI） and Normalized Ratio Chlorophyll Index （PSND）, were correlated with the chlorophyll of sand sagebrush, and the correlation coefficients had reached over 0.9 at 99% confidence level. The determination coefficient was R SIPI 〉 R PSND 〉 R PSSR 〉 R NDVI2 〉 R NDVI1 〉 R SR in the estimation model establish by six planting index, that is, the coefficient of determination was above 0.9, and the highest determination coefficient of spectral index SIPI had reached 0.983. It showed that the model established by SIPI could play a good role in prediction and it was suitable for estimating the chlorophyll of sand sagebrush.