IEEE Signal Processing Society Colloquia

High-Resolution Radar Imaging With Applications to Astronomy

David Munson, Jr.
University of Illinois at Urbana-Champaign

Computed imaging systems are capable of synthesizing imagery having extraordinary resolution, by processing data collected from an array of sensors or from a single sensor moved to different spatial locations. Many of these systems, such as synthetic aperture radar (SAR), interferometric radio astronomy, computer tomography, magnetic resonance imaging, electron microscopy, and x-ray crystallography, acquire partial Fourier data, and then employ sophisticated digital processing techniques to form 2-D, 3-D, and 4-D images.

After providing an overview of Fourier-based computed imaging, this talk will focus on SAR applied to radar astronomy. In so doing, we will point out relationships between three computed imaging systems: range-Doppler radar astronomy, computer tomography, and SAR. We will begin by reviewing the range-Doppler approach to planetary radar imaging. We will then explore the possibility of using a polar-format SAR model instead. This latter model is based on theorems from computer tomography, and avoids the traditional nemesis of motion through resolution cells. Using Lunar data from Arecibo Observatory, we demonstrate that polar-format SAR processing can provide significantly improved resolution in radar astronomy.

When imaging 3-D objects, such as asteroids, which have irregular shapes, the polar-format SAR method requires collection of a huge, and impractical, amount of data. To circumvent this difficulty, we propose a new method for radar imaging of 3-D surfaces, which casts the imaging problem into the form of a high-resolution spectral estimation problem. Promising results are shown using simulated 3-D data.