Linear Mixture Modeling Approach for Estimating Cotton Canopy Ground Cover using Satellite Multispectral Imagery

Researchers have sought techniques for estimating plant canopy characteristics from remote sensing data that are not site-specific. In this study, a procedure for estimating cotton (Gossypium spp.) canopy ground cover based on a linear mixture modeling approach was tested using Landsat multispectral imagery. Ground cover estimated from scene reflectance in the red and near-infrared wavebands was compared with corresponding field measurements. Statistical analysis of the results suggest that the procedure produced accurate, unbiased estimates of cotton canopy ground cover. The procedure appeared to implicitly compensate for the effects of shadows in the scene. These results agree with the findings of earlier studies and support the proposition that this procedure should be more universally applicable than commonly used empirical curve-fit techniques.