Estimation of maize canopy properties from remote sensing by inversion of 1-D and 4-D models

Abstract

The inversion of canopy reflectance models is widely used for the retrieval of vegetation properties from remote sensing. However the accuracy of the estimates depends on a range of factors, most notably the realism with which the canopy is represented by the models and the possibility of introducing a priori knowledge on canopy characteristics to constrain the inversion procedure. The objective of the present work was to compare the performances and operational limitations of two contrasting types of radiative transfer models: a classical one-dimensional canopy reflectance model, PROSPECT+SAIL (PROSAIL), and a three-dimensional dynamic (4-D) maize model. The latter introduces greater realism into the description of the canopy structure and implicit a priori information on the crop. The assessment was carried out with multiple view angle data recorded from field experiments on maize at stages V5 to V8. The simplex numerical optimization algorithm was used to invert the two models, using spectral reflectance data for PROSAIL and gap fraction data for the 4-D maize model. Leaf area index (LAI) was estimated with a RMSE of 0.48 for PROSAIL and 0.35 for the 4-D model. Retrieval of average leaf inclination angle (ALA) was problematic with both models. The effect of the number and distribution of observation view angles was examined, and the results highlight the advantage of oblique angle measurements.