Spatially Varying Aperture Weighting for Sidelobe Reduction and Resolution Enhancement of Imagery
Abstract
Synthetic aperture radar (SAR) engineers continue to search for algorithms and techniques that reduce sidelobes and improve the resolution of a given data set. The classical estimation of the power spectrum of a signal from a finite record of data is typically made using a discrete Fourier transform (DFT). To mitigate errors due to the data record being finite, one typically applies various windows (or apodizations) before transformation. These windows fall into two categories: raised cosine- based (e.g., Harming, Hannning, Blackman-Harris, and Nutall windows), and noncosine-based (e.g., Dolph-Chebyshev, Taylor, Kaiser, and Gaussian windows). These windows trade off resolving power for improved (reduced) sidelobes. Similarly, SAR imagery exhibits sidelobes because the images are derived from a finite aperture. The same windows are applied to SAR data to reduce the sidelobes. This report describes a method of allowing a raised cosine-based window to be adaptively changed at each output data point (or pixel). The technique is spatially variant apodization (SVA). This spatially adaptive window maintains the resolution normally associated with rectangular weighting. However, it simultaneously reduces the sidelobes commensurate with the order of the filter (i.e., the number of cosine terms used). Results are shown on imagery from the U.S. Army Research Laboratory's ultra-wideband Boom-SAR system. It is important to note that SVA is not limited to SAR applications. It is applied in the image domain and is applicable to all systems that produce images.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1998
- Accession Number
- ADA355541
Entities
People
- John W. Mccorkle
- Matthew R Bennett
Organizations
- United States Army Research Laboratory