New Algorithms and Sparse Regularization for Synthetic Aperture Radar Imaging
Abstract
The PI and his collaborators proposed an algorithm to form a synthetic aperture radar (SAR) image in low algorithmic complexity. It is based on the so-called butterfly scheme. Control over the accuracy is provided. Speedups in the hundreds are reported on the Air Force's GOTCHA dataset. The PI and his collaborators also investigated the possibility of forming super-resolved images from bandlimited pulse-echo data with ideas of sparse optimization that bear a link to compressed sensing. The difficulty of super-resolution is summarized in a single number, a principal angle between subspaces, which also governs the algorithmic complexity of the minimization.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 2013
- Accession Number
- ADA580544
Entities
People
- Laurent Demanet
Organizations
- Massachusetts Institute of Technology