|Stochastic Systems Group|
Wide-Angle SAR Imaging
The first half of the talk will be presented at SPIE Defense and Security Symposium, Algorithms for Synthetic Aperture Radar XIII, April 18, 2006. We consider the problem of jointly forming images and characterizing anisotropy from wide-angle synthetic aperture radar (SAR) measurements. We present a method based on a sparse representation of aspect-dependent scattering with an overcomplete basis composed of basis vectors with varying levels of angular persistence, enforcing sparsity through regularization based on the l_k-norm, k < 1. This formulation leads to an optimization problem that is solved through a robust quasi-Newton method. We also develop a graph-structured interpretation of the overcomplete basis leading towards greedy algorithms using guided depth-first search with appropriate stopping conditions and search heuristics.
In the second half, two extensions to the anisotropy characterization problem are presented. The first deals with an effect that is prominent in wide-angle imaging much more so than in narrow-angle imaging --- that certain scattering mechanisms migrate as a function of aspect angle. That is to say that a scattering center appears in different spatial locations depending on the viewing direction. Migratory scattering centers, or 'movers' have not been given much heed in the past. We build upon our previous work on jointly forming images and characterizing anisotropy to incorporate migratory scattering centers.
The Hough transform has traditionally been used on already formed binary images for the detection of parameterized shapes. As an extension to our sparsifying regularization formulation, we reverse this paradigm by augmenting sparsifying regularization with regularization terms operating in Hough space to incorporate prior knowledge and favor or suppress certain spatial interrelationships among scattering centers within the image formation process.
Problems with this site should be emailed to email@example.com