Comparison of three radiative transfer model inversion techniques to estimate canopy biophysical variables from remote sensing data

The objective of this study was to compare three model inversion techniques (neural networks (NNT), look up tables (LUT), iterative optimization (OPT)) for the estimation of four main biophysical variables: LAI, C_ab, f_cover and f_PAR. The canopy radiative transfer model used is SAIL coupled with PROSPECT leaf model. Results show that OPT performs better when applied to datasets consistent with the radiative transfer model assumptions used in the inversion approach. Conversely, NNT shows significant improvement when applied to different datasets. The authors observed also large differences in the retrieval performances between the biophysical variables investigated, f_PAR and f_cover being better estimated than LAI and C_ab