Linear mixture modeling applied to remote sensing monitoring of winter wheat areas

Moderate Resolution Imaging Spectroradiometer (MODIS) has advantages in the following aspects: multi-spectra, multi-temporal and freely obtained. Using traditional classification methods, super-classification and vegetation index thresholds, based on MODIS data during the winter wheat regreening stage, the distribution of winter wheat area was investigated in this study. Meanwhile, aiming at the characteristics that most remote sensing pixels are mixed pixels, application of linear mixture modeling to unmixing the planting area of winter wheat was mainly studied. Comparing the precision of different classification methods for the planting area of winter wheat, using linear mixture modeling, the major part (98.45%) of the root-mean-square error is smaller than 0.01, the relative error is approximately 3% compared with the actual winter wheat field data, obviously superior to the precision of traditional remote sensing classification method.