Gaussian Process Regression Model for Crop Biophysical Parameter Retrieval from Multi-Polarized C-Band SAR Data

Biophysical parameter retrieval using remote sensing has long been utilized for crop yield forecasting and economic practices. Remote can provide information across a large spatial extent in timely manner within season. Plant Area Index (PAI), Vegetation Water Content (VWC), Wet-Biomass (WB) play vital role estimating growth helping farmers make market decisions. Many parametric non-parametric machine learning techniques have to estimate these parameters. A general approach that follows Bayesian framework is the Gaussian Process (GP). The parameters of this process-based technique are assumed be random variables with joint distribution. purpose work investigate Regression (GPR) models retrieve biophysical three annual crops utilizing combinations multiple polarizations from C-band SAR data. RADARSAT-2 full-polarimetric images situ measurements wheat, canola, soybeans obtained SMAPVEX16 campaign over Manitoba, Canada, used evaluate performance GPR models. results research demonstrate both full-pol (HH+HV+VV) combination dual-pol (HV+VV) configuration PAI, VWC, WB crops.