Clutter statistics along edge features in synthetic aperture radar imagery

Abstract

The K-distribution model for clutter over distributed targets in Synthetic Aperture Radar (SAR) images is extended to include clutter along edges and thin linear targets. The physical mechanisms generating edge features are analyzed. Assuming a boxcar system response, the detected field from a resolution cell along an edge is shown to be expressible in terms of sums of random contributions from different scatterer populations, plus single phasors representing specular and secondary scattering contributions. The clutter statistics resulting from this expression are derived assuming that speckle is fully developed and that the number of scatterers from each population is controlled by a birth-death-immigration process. In the absence of specular and secondary scattering the detected field accords with the product model for radar imaging, i.e. the field is given by the product of a complex circularly symmetric zero-mean Gaussian speckle process and the square-root of the imaged surface cross-section (SCS). The imaged SCS is given by a weighted sum of gamma distributed random variables, each one of which represents the SCS due to a single distributed target. Under certain limiting conditions the imaged SCS is gamma distributed or constant. Some of the results are confirmed by statistical measurements taken along edges and thin linear targets.