Feature-Based Methods for Landmine Detection with Ground Penetrating Radar

Abstract

The subject research was performed at the University of Florida between December 2005 and December 2008. The research was performed to support the ability to detect landmines in an automated fashion using ground-penetrating radar (GPR) array sensors employed in systems being studied by NVESD. The work was concerned with discovering and evaluating i) different types of features that, when extracted from signals associated with GPR signals captured over regions of earth, can help one identify the presence or absence of landmines and landmine-like objects; ii) algorithms and techniques that can employ these features to distinguish between landmines and non-mines; and iii) fuse the results of multiple discriminators to yield improved discrimination performance.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 27, 2012
Accession Number
ADA569913

Entities

People

  • Jeremy Bolton
  • Joseph N. Wilson
  • Paul D. Gader

Organizations

  • University of Florida

Tags

Communities of Interest

  • C4I
  • Counter IED
  • Energy and Power Technologies
  • Human Systems
  • Materials and Manufacturing Processes
  • Sensors

DTIC Thesaurus Topics

  • Anti-Personnel Mines
  • Anti-Tank Mines
  • Artificial Intelligence
  • Bayesian Networks
  • Computational Science
  • Computer Vision
  • Data Mining
  • Detectors
  • Information Processing
  • Information Science
  • Machine Learning
  • Mathematical Filters
  • Network Science
  • Neural Networks
  • Pattern Recognition
  • Three Dimensional
  • Two Dimensional

Readers

  • Munitions and Ordnance Engineering
  • Neural Network Machine Learning.
  • Radar Systems Engineering.