Feature Extraction and Classification of Magnetic and EMI Data, San Luis Obispo, CA

Abstract

An Unexploded Ordnance discrimination study was conducted at the Former Camp San Luis Obispo (SLO), California. The objective was to discriminate potentially hazardous 2.36" rockets and 60 mm, 81 mm and 4.2" caliber mortars from non-hazardous shrapnel, range and cultural debris. In this report, we describe the performance of twelve different discrimination techniques that utilized data from a number of sensors deployed in full coverage and cued interrogation mode. Two important conclusions can be drawn from the results presented here. Firstly, by appropriate use of discrimination metrics applied to production quality EM-61 data, it is possible to significantly reduce the number of clutter items excavated without missing any targets of interest. Secondly, the next generation of EM sensors, when deployed in a cued-interrogation mode, result in significant additional reductions in the number of clutter items excavated. Furthermore, the next generation sensors can usually distinguish different UXO types from one another.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 2010
Accession Number
ADA579039

Entities

People

  • Douglas Oldenburg
  • Jon Jacobson
  • Kevin Kingdon
  • Laurens Beran
  • Len Pasion
  • Lin P. Song
  • Nicolas Lhomme
  • Stephen Billings

Organizations

  • Sky Research

Tags

Communities of Interest

  • Weapons Technologies

DTIC Thesaurus Topics

  • Computational Science
  • Data Analysis
  • Data Mining
  • Delphi Method
  • Detection
  • Detectors
  • Dimensionality Reduction
  • Feature Extraction
  • Information Science
  • Kernel Functions
  • Machine Learning
  • Ordnance Locators
  • Pattern Recognition
  • Supervised Machine Learning
  • Two Dimensional
  • Unsupervised Machine Learning
  • Uxo Detection

Readers

  • Computational Modeling and Simulation
  • Military/Explosive Ordnance Disposal (EOD) Technology

Technology Areas

  • AI & ML