AI (Artificial Intelligence) Analysis for Automatic Target Recognition,

Abstract

Research directed toward the goal of integrating contextual scene information into the automatic target classification process is described. The basic approach adopted is an information fusion and feedback architecture centered around an artificial intelligence (AI) production system. A parallel organization of specialized algorithms extracts complementary contextual discriminants from the local, global, and temporal image data. A high-level relation data structure integrates this imagery information with non-imagery scene data, such as a priori knowledge of terrain, environment, and expected threat, thereby providing a symbolic representation of the total dynamic scene. Extensive domain-specific and general knowledge applied to the scene representation by an AI Expert System exploits the associative evidence to deduce additional facts, infer signatures, and perform collective target class decisions. In addition, an AI global/local control strategy routes derived knowledge and revised hypotheses to all extraction algorithms to bootstrap overall preprocessor intelligence and provide adaptive optimization through feedback. (Author)

Document Details

Document Type
Technical Report
Publication Date
Jan 31, 1984
Accession Number
ADP003023

Entities

People

  • A. J. Spiessbach
  • J. F. Gilmore

Organizations

  • Georgia Tech

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Artificial Intelligence
  • Automatic
  • Classification
  • Expert Systems
  • Feedback
  • Recognition
  • Target Classification
  • Target Recognition

Fields of Study

  • Computer science

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Artificial Intelligence
  • Computer Vision.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy