Li Xiaobing¹, Long Huiling², Wang Hong¹

¹ State Key Laboratory of Earth Surface Processes and Resource Ecology, College of Resources Science and Technology, Beijing Normal University, Beijing, 100875, China;

² Beijing Research Center for Information Technology in Agriculture, Beijing Academy of Agriculture and Forestry Sciences, Beijing, 100097, China.

 

Abstract: Changes of key parameters of vegetation are essential indicators of ecosystem and global change. Hyperspectral data, as a powerful tool to estimate vegetation parameters, needs to be used more efficiently and effectively, especially in the aspect of massive information extraction. The objectives of the present study were to provide guidance on how to select the optimal subset of hyperspectral data to improve the accuracy of estimating vegetation cover using hyperspectral data measured in the field, and to compare the predictive ability of several estimation models. Based on the field-measured hyperspectral curves for completely covered land, bare soil, and the vegetation canopy, we used vegetation cover data obtained by analyzing digital camera photos and different vegetation indices to calculate the accuracy of estimation of vegetation cover by the different models and we discuss differences among the models. We found the most accurate estimate of vegetation cover in our study area using a single optimal combination of wavelengths based on MSAVI2 indices and the semi-empirical model proposed by Gutman and Ignatov.

 

Keywords: Vegetation cover; Hyperspectrum; Field measurement; Wavelength selection.