Estimation and Mapping of crop biomass and height

To estimate crop AGB and height, multiple modeling approaches were implemented using S-1 polarizations (VV, VH, VH+VV, and VH-VV), the proposed SAR texture indices (RSTI and NDSTI), and six S-2 VIs. Univariate regression models were developed using five commonly algorithms: linear, polynomial, exponential, power, and logarithmic regression. These models were used to assess the individual predictive power of each input feature. In addition, bivariate models were constructed using partial least squares regression (PLSR) and GPR to integrate SAR texture indices and VIs. These models were designed to evaluate the synergistic potential of combining SAR and optical features, particularly for alleviating the saturation effect often encountered at medium to high biomass levels.