叶片化学组分成像光谱遥感探测机理分析

利用地面光谱仪的测量数据 ,进行了成像光谱遥感探测叶片化学组分的机理性研究。采用多元逐步回归方法 ,分析了鲜叶片 7种化学组分含量与其光谱特性的统计关系 ,分别建立了反射率 ρ及其变化式 1/ρ、logρ和ρ的一阶导数Kρ 与化学组分含量的统计方程 ,并对这 4个指标的性能进行了比较和评价。结果表明 ,在 95 %的置信水平下 ,可以由叶片的精细光谱特征较好地反映出化学组分含量 ;特别是利用Kρ 作为因子 ,使置信水平提高到 99% ,尤以对粗蛋白质、N、K含量反映最好 ,R2 均达到 0 8以上 ,粗蛋白质可达 0 95 6 4,从而为进一步探讨在中国利用成像光谱遥感探测叶片化学组分奠定了基础

Mechanism Analysis of Leaf Biochemical Concentration by High Spectral Remote Sensing

This paper presents the mechanism research on predicting the biochemical concentration of fresh leaves by high spectral remote sensing. Based on analyzing the concentrations of seven chemical components, including total chlorophyll, water, crude protein, soluble sugar, N, P and K, with certain chemical methods and detecting their optical properties with surface spectrometre, we establish the statistical relationships between the concentration and reflectance through the stepwise multiple regression method. So did the relationships between the concentrations and several transformations of reflectance such as the reciprocal, the logarithm, and the first derivative of the reflectance. The results show good prediction performance for chlorophyll, water, crude protein, N and K with high values of the squared multiple correlation coefficients ( R 2) and high confidence level (>95%). Especially, R 2 value of the corralation between crude protein concentration and the first derivative of reflectance is 0.9564, which is the best result in the study of the fresh leaf's biochemistry. The research lays a good basis for further discussion on predicting leaf biochemical concentration by high spectral remote sensing in China.