The Sources of Variability in the Search Process

Abstract

Modelling of camouflage concealment and detection needs to consider the terrain in which a target will appear. A variety of capabilities for evaluating target signatures through to the human response now exists. The modelling approach adopted at British Aerospace research centre for many years has been a statistical one. Although image analysis techniques have been explored it has been cost effective for our purposes to stay with the statistical model ORACLE. A complex problem in modelling human visual performance is to find an adequate relationship between recognition thresholds across the visual field and simple target descriptions. The ORACLE model represents recognition as the resolution of a fraction of the target perimeter. The fractional perimeter concept has been applied further to representing average observer performance in structured scenes. Modelling of the SEARCH 2 data2 was found to need a similar distribution of fractional perimeter values to a previous UK field trial. The use of a statistical lobe model for analysing search and recognition can be supported for generic information.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 01, 2000
Accession Number
ADP010544

Tags

Communities of Interest

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

DTIC Thesaurus Topics

  • Acquisition
  • Aspect Ratio
  • Character Recognition
  • Contrast
  • Cost Effectiveness
  • Data Sets
  • Detection
  • Digital Data
  • Eye Movements
  • Identification
  • Images
  • Peripheral Vision
  • Recognition
  • Statistics
  • Target Acquisition
  • Target Signatures
  • Targets

Readers

  • Sensor Fusion and Tracking Systems.
  • Theoretical Analysis.
  • Vision Science/Vision Psychology/Cognitive Neuroscience.

Technology Areas

  • Space
  • Space - Space Objects