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.