The Rayleigh Generalized Autoregressive Score Model for SAR Data Interpretation

This letter introduces the Rayleigh generalized autoregressive score (Ray-GAS) model, a dynamic model useful for synthetic aperture radar (SAR) data interpretation. It is derived from the generalized autoregressive score (GAS) framework, assuming that the Rayleigh conditional mean is an index-varying parameter. We discuss estimation, diagnostic, and prediction tools. Additionally, we perform numerical experiments with simulated and measured single-look amplitude SAR data for forest and lake region data. The results show the Ray-GAS model’s benefits in understanding stochastic behavior and filtering SAR amplitude returns.