The Shift and Scale Invariant Fourier-Mellin Transform for Radar Applications

Abstract

Automatic classification of targets viewed by radar is complicated by variations in target aspect relative to the radar line-of-sight (RLOS). This report investigates the possibility of reducing the effects of target aspect by using the scale invariance of the Mellin transform. The properties of the Mellin transform are developed in analogy with the Fourier transform and illustrated using simple test functions and digitally implemented transforms. Simulated radar ship profiles demonstrate that a change in aspect is not equivalent to a change in target scale for realistic targets, however. Automatic classification results, for both simulated and actual radar ship profiles, confirm that using a combination Fourier-Mellin transform for feature selection appears at best comparable to the results obtained using the Fourier transform alone for feature selection.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 14, 1980
Accession Number
ADA093690

Entities

People

  • Lyle H. Johnson

Organizations

  • Massachusetts Institute of Technology

Tags

Communities of Interest

  • Ground and Sea Platforms
  • Materials and Manufacturing Processes
  • Sensors
  • Weapons Technologies

DTIC Thesaurus Topics

  • Accuracy
  • Aspect Angle
  • Automatic
  • Classification
  • Discrete Fourier Transforms
  • Feature Selection
  • Frequency
  • Frequency Domain
  • Identification
  • Integrals
  • Invariance
  • Line Of Sight
  • Pattern Recognition
  • Recognition
  • Slant Range
  • Three Dimensional
  • United States

Readers

  • Approximation Theory.
  • Computer Vision.
  • Radar Systems Engineering.