Prediction of the Recognition of Real Objects as a Function of Photometric and Geometric Characteristics

Abstract

The purpose of this research was to predict target-by-target acquisition performance of air-to-ground imagery from microdensitometrically determined photometric and geometric characteristics of the scene. Results showed that it is feasible to predict the ground range at which a given target will be detected by an airborne observer. This prediction could be made totally automatically, given reconnaissance imagery, a microdensitometer, and a small computer. Seventeen characteristics of targets, backgrounds, and target/background relationships which reliably correlate to target acquisition performance were identified. Sixty regression equations combining these variables into linear predictive models of target acquisition performance were developed from one set of targets and mission conditions and cross-validated against targets contained in three different reconnaissance missions.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1978
Accession Number
ADA071118

Entities

People

  • Deborah G. Bonnet
  • Harry L. Snyder

Organizations

  • Virginia Tech

Tags

Communities of Interest

  • Biomedical
  • Energy and Power Technologies
  • Ground and Sea Platforms
  • Human Systems
  • Sensors

DTIC Thesaurus Topics

  • Cameras
  • Computer Programs
  • Computers
  • Data Analysis
  • Data Mining
  • Data Science
  • Databases
  • Factor Analysis
  • Information Science
  • Magnetic Tape
  • Measurement
  • Photographs
  • Predictive Modeling
  • Regression Analysis
  • Social Sciences
  • Target Acquisition
  • Target Detection

Fields of Study

  • Physics

Readers

  • Computational Modeling and Simulation
  • Geodesy
  • Image Processing and Computer Vision.