Identification of Radar Targets by Pattern Recognition.
Abstract
The problem of identifying radar targets without predetermined knowledge or measurements of their aspect angle is examined. The proposed solution is to use radar magnitude and phase versus frequency data in the resonance region as input data for pattern recognition techniques. Practical methods are developed for measuring radar phase and dealing with polarization effects. Pattern recognition algorithms that optimize separation between pairs of input data are developed. These algorithms are based on digital spatial frequency filtering of curves of the radar return versus the radar frequency. The technique is simulated on a digital computer for three objects - a cylinder, a cone, and a step cylinder - all of about the same size. Results show that at most aspect angles, satisfactory recognition is attained at moderate signal-to-noise ratios. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 26, 1973
- Accession Number
- AD0764713
Entities
People
- William B. Goggins Jr.
Organizations
- Air Force Institute of Technology