Soil moisture retrieval using SAR data and a priori roughness information

In this paper, two model-based methods for soil moisture retrieval from SAR measurements are described. The performances of the two methods are compared using, simulated L-band data obtained by exploiting the integral equation model (IEM). The first method is based on a previously developed soil moisture retrieval algorithm. It estimates the soil moisture content by approximating the IEM inverse function using a neural network approach. The second method is an iterative technique which adopts a searching strategy by using the IEM direct model. The two methods are also compared when using a priori information, such as the roughness state of the bare soil surfaces