A Method of Analyzing Air System Performance Based on Shape Distortion

Abstract

This research developed a measure, we call the Maximum Resolvable Polygon (MRP), which estimates the degree of shape distortion introduced by an imaging system. The MRP relies on a computer simulation of the imaging system. While such a computer simulation requires more extensive computer power than the computation of an analytical formula like the MRT, the MRP may still be considered relatively easy to compute. All software simulations and experiments reported here were performed on a personal computer. A new test pattern set has been proposed to replace the bar pattern which is used to analyze the performance of man-in-the-loop systems. Target-like test patterns rely on the use of regular polygons and are parameterized by the number of sides in the polygons. Clutter-like objects are also derived from this test pattern set. A theoretical analysis of the MRP indicates that it estimates the target shape conditions at which a target recognition system will produce a 50% error rate. Simulation experiments have compared the performance which the MRP measure predicts that an automatic target recognition system would have with the performance of the human visual pattern recognition system. (jhd)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 04, 1990
Accession Number
ADA226193

Entities

People

  • Michael S. Mort

Tags

Communities of Interest

  • Energy and Power Technologies
  • Materials and Manufacturing Processes
  • Sensors

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Cartesian Coordinates
  • Computational Science
  • Computations
  • Computer Programs
  • Computer Simulations
  • Computers
  • Detection
  • Detectors
  • Distortion
  • Geometry
  • Image Processing
  • Information Processing
  • Pattern Recognition
  • Probability
  • Target Recognition
  • Two Dimensional

Fields of Study

  • Engineering

Readers

  • Image Processing and Computer Vision.
  • Mathematical Modeling and Probability Theory.
  • Regression Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference