Abstract: Fractional green vegetation coverage (FC) is the most effective indicator of land desertification, and remote sensing is an important means to obtain regional scale vegetation coverage. The methods of processing EO-1 Hyperion hyperspectral image and estimating vegetation coverage using quantitative remote sensing are researched in this paper. We compare two different methods of estimating vegetation coverage from EO-1 Hyperion data. The first method is based on Li-Strahler Geometric-Optical model and Spectral Mixture Analysis(SMA) technique. The second method is based on mixed-pixel models. Results of vegetation coverage by the two methods are compared, and are further verified by measured data in the experimental field of Helan Mountain. The results indicate that the Li-Strahler Geometric-Optical model inversion (R2=0.76, RMSE=0.06) performs better than the mixed-pixel model inversion (R2=0.71, RMSE=0.07) for FC retrieval.

Key words: hyperspectral remote sensing, EO-1 Hyperion image, vegetation coverage, geometric optical model, mixed pixel model