基于高光谱数据的杨树叶片干物质含量的估算

将杨树叶片实测的理化数据和土壤背景光谱数据作为PROSPECT和PROSAIL模型的输入参数,输出杨树叶片高光谱数据模拟值,通过实测获得叶片尺度和冠层尺度干物质含量、等效水厚度以及高光谱数据,利用统计方法,分别对两种尺度杨树叶片干物质含量进行分析。结果表明:基于归一化指数计算方法,杨树叶片尺度和冠层尺度的最佳估算干物质含量的干物质含量归一化指数(NDMI)波段组合分别为1685,1704nm和1551,2143nm,使用偏最小二乘法分别对筛选波段组成的NDMI(1685,1704)指数和NDMI(1551,2143)指数构建叶片尺度干物质含量和冠层尺度干物质含量的估算模型,叶片尺度干物质含量估算模型精度为R2=0.663,RMSE=0.001g·cm-2,冠层尺度精度为R2=0.91,RMSE=16.7g·m-2。可见,高光谱技术对杨树叶片干物质含量的估算具有较高的精度,可为杨树叶片干物质含量的快速、无损估算提供参考依据。

Estimating Dry Matter Content of Poplar Leaf by Hyperspectral Data

In order to explore estimating the dry matter content of poplar leaves accurately,rapidly and accurately and nondestructive with hyperspectral data.Simulated values of poplar leave hyperspectral data which were leaf scales and canopy scales were obtained by PROSPECT and PROSAIL models which input parameters were measured by the physical and chemical data and soil background spectral data.At the same time,the dry matter content,equivalent water thickness and hyperspectral data were obtained from the measured leaf scales and canopy scales.Finally,the statistical methods were used to analyze the dry matter content in two scales of poplar leaves.The results showed that,the best band combination of the normalized difference vegetation index of dry matter was respectively band combination between 1685nm and 1704nm and band combination between 155 l nm and 2143nm,respectively in leaf scales and canopy scales.Dry matter content models of leaf and canopy scales were built respectively by NDMI (1685,1704) and NDMI(1551,2143).The model validation accuracy of the leaf scale dry matter content was R2=0.663,RMSE=0.001g.cm-2,and the model validation accuracy of the canopy scale dry matter content was R2=0.91,RMSE=16.7g.m-2.Obviously,the hyperspectral technique has a high precision for estimating the dry matter content of poplar leaves,which can provide reference for the rapid and nondestructive estimation of dry matter content of poplar leaves.