Synthetic aperture radar image of agricultural fields with surface drainage network: simulation and spatial information retrieval

Abstract

We develop a 3-D model to simulate the synthetic aperture radar (SAR) image formation process of an undulated vegetation canopy such as corn grown in fields with large periodic drainage reliefs. We explain how the simulated SAR image of undulated vegetation medium is obtained by the convolution of a 2-D slice of the 3-D simulated SAR system point spread function [(PSF), emulating the SAR beam modeled by a cosine modulated Gaussian], with the 2-D projection of the observed undulated vegetation canopy (modeled with scatterers randomly distributed in 3-D undulated space) followed by the extraction of each look envelope, the summation of looks, and sampling in azimuthal and range directions. Our model is useful to study the parameters involved in the formation and the analysis of SAR images of undulated vegetation medium. Validation of simulations made with actual SAR images shows that undulated corn crop canopies are well characterized by the mean contrast of the backscattered amplitude, the shape parameter ? of the K distribution, the coefficient of variation, and the autocorrelation function of the backscattered amplitude. These experiments also show that the discrimination of corn parcels with or without subsurface drainage and the monitoring of the corn plant height differential across the drainage ridge during the growing season are possible.