论文部分内容阅读
类胡萝卜素(Car)作为植物主要色素,对诊断植被生理状态有重要作用。于2013年4月和7月采集闽江口秋茄(Kandelia candel)叶片,室内测定其叶片正面和背面反射光谱,同时测定其Car含量[单位面积(μg·cm-2)和单位质量(mg·g-1)]。选取常见Car含量估算的光谱参数,同时分析确定最佳比值植被指数(SR),基于回归分析,建立秋茄叶片Car含量估算与验证模型。结果表明,叶片光谱反射率表现为叶片背面大于正面(350~2350 nm);基于叶面背面光谱计算的SR与叶片Car含量(μg·cm-2)的相关系数优于其他组合,相关系数较高的区域分布在520~540 nm与1000~1100 nm波段组合,700~720 nm与800~1100 nm波段组合;基于背面光谱计算的大部分光谱参数与Car含量(μg·cm-2)的相关系数要高于基于正面光谱计算的。因此,以叶片背面光谱作为Car含量估算的光谱数据,以单位面积Car含量为估算量纲建立反演模型。本研究表明,光谱指数LCI、DD、NDVI(770,713)、NDVI(773,562)、SR(723,770)和SR(1000,700)均可实现Car含量的反演,估算与检验模型的R2均>0.65,RMSE均<1.52;并且新构建的SR(1000,700)估算精度最好,模型和检验R2分别为0.77和0.87,模型和检验RMSE分别为1.08和1.11。这些预味着基于高光谱遥感对闽江河口湿地秋茄Car含量进行估算是可行的。
Carotenoids (Car) as the main pigment of plants play an important role in the diagnosis of physiological status of vegetation. The leaves of Kandelia candel were collected in April and July 2013, and the leaf front and back reflectance spectra were measured indoors. The content of Car [unit area (μg · cm -2) and unit mass mg · g-1) ]. Spectral parameters estimated by common Car content were selected, and the optimal ratio of vegetation index (SR) was determined. Based on regression analysis, the Carbazole content in leaves of candelia was estimated and validated. The results showed that the spectral reflectance of leaves was larger than that of the front (350 ~ 2350 nm). The correlation coefficient between SR and leaf Car content (μg · cm-2) was better than other combinations based on the leaf back surface spectra, the correlation coefficient was High region distributions are combined in the 520-540 nm and 1000-1100 nm bands and in the 700-720 nm and 800-1100 nm bands. Most of the spectral parameters calculated based on back surface spectra are related to Car content (μg · cm -2) The coefficient is higher than calculated based on the front spectrum. Therefore, the leaf back spectrum was used as the estimated spectral data of Car content, and the unit area Car content was used as the estimation dimension to establish the inversion model. This study shows that the inversion of Car content can be achieved with the spectral indices of LCI, DD, NDVI (770,713), NDVI (773,562), SR (723,770) and SR (1000,700) RMSE <1.52; and the newly constructed SR (1000,700) has the best estimation accuracy with the model and test R2 of 0.77 and 0.87, respectively, and the model and test RMSE of 1.08 and 1.11, respectively. It is feasible to estimate Car content in the candel based on hyperspectral remote sensing in the Min River Estuary wetlands.