Target identification by means of low frequency radar returns
Abstract
A target identification technique is presented. It is based on low frequency radar returns. A large variety of objects is shown to be reliably classified in the presence of substantial amounts of noise. The target complexity varies form cylinders and cubes to modern combat airplanes. Two classification algorithms are used and compared, a linear discriminant method and a nearest neighbor rule. Both multiclass and pairwise classifications are carried out and extensively illustrated. The majority of results presented involve only the amplitudes of the radar returns, but the last part of the report describes the very impressive effects of phase on classification performance.
- Publication:
-
NASA STI/Recon Technical Report N
- Pub Date:
- May 1974
- Bibcode:
- 1974STIN...7510293L
- Keywords:
-
- Low Frequencies;
- Radar Echoes;
- Target Recognition;
- Backscattering;
- Data Processing;
- Radar Cross Sections;
- Radar Targets;
- Communications and Radar