Hyperspectral remote sensing estimation models for aboveground fresh biomass in Gannan grassland

In order to promote the application of hyperspectral remote sensing in the dynamic monitoring and yield estimation of grassland,the canopy spectral reflectance and the aboveground fresh biomass corresponding to the spectra of natural grassland were measured in Gannan grassland.This paper analyzed the spectral reflectance characteristics of four main grassland types,the correlation between the aboveground fresh biomass and reflective spectrum,and the correlation between aboveground fresh biomass and the first derivative spectrum.Using characteristic bands and their combination that were strongly correlated to the aboveground fresh biomass,this paper defined hyperspectral parameters as variables.Thus,the hyperspectral remote sensing estimation models of the grass aboveground fresh biomass were established in Gannan prairie.The estimation models were tested by the experiment data.The results showed that estimation model of D723 [y=3.526 lnD723+18.923] was the best,and it's RMSE,relative error,and the correlation coefficient between the estimated value and measured value were 0.208 3,8.8% and 0.896,Therefore,the model could preferably estimate the grass aboveground fresh biomass in Gannan prairie.