The Use of Classifiers and Microprocessors for Target Identification Using Millimeter Wave Signatures.
Abstract
The problem of implementing three basic classification techniques for target acquisition in millimeter wave guided weapon systems is discussed in this report. The three classifiers are: (1) the maximum likelihood classifier, (2) the nearest neighbor classifier, and (3) the linear classifier. The mathematics of each classifier is explicitly delineated in order to assess memory storage and computation load requirements. Using these results, the applicability of microprocessors for implementing these classifiers into fieldable systems is considered. It is shown that the classifiers are easily adaptable to experimental target signatures and that the use of microprocessors is requisite. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 26, 1979
- Accession Number
- ADA074840
Entities
People
- P. Martin Alexander