A Radar Target Data Processing and Identification System

摘要

A system is described which is capable of testing the identifiability of various targets based on their multifrequency radar returns. Three software packages designed to manipulate, analyze and identify multivariate data are described. In the first, called RADSYS, data may be input to the computer and manipulated in a number of ways such as interpolation, dimensionality reduction by data component selection, normalization by a multiplicative constant, etc. RADSYS also contains a recognition routine which uses a nearest neighbor rule for the identification of data sets. The second package, called DAD, is used to project the n dimensional data sets onto a two dimensional plane and display this data on a CRT. The final program, called MST, is used to compute the minimal spanning tree of arbitrary sets of data. (Author)