Grazing angle dependency of SAR imagery

A problem is associated with the repeatability of the radar geometry at which the two scenes are imaged. If there is a significant difference between the two geometries, then the scattering properties between the two views will be different resulting in an increased probability of false detection (of changes). This paper examines the extent to which this occurs by analysing the same scene viewed at a series of grazing angles ranging from 1° to 10°. The authors briefly describe the processing techniques used to provide high quality imagery and the change detection technique itself. They describe the experimentation performed to enable the effects of small grazing angle changes to be evaluated, and present the results of experimental observations. The implications of these for the robustness of applications such as change detection are discussed