Electromagnetic Models for UXO Detection and Classification in Permeable Soils

Abstract

Support Vector Machine (SVM) and neural networks (NN), are applied to classifying metallic objects according to size using the expansion coefficients of their magneto-quasistatic (MQS) response in the spheroidal coordinate system. The classified objects include homogeneous spheroids and composite metallic assemblages meant to resemble unexploded ordnance. An analytical model is used to generate the necessary training data for each learning method. SVM and NN are shown to be successful in classifying three different types of objects on the basis of size. They are capable of fast classification, making them suitable for real-time application. Furthermore both methods are robust and have a good tolerance of 20 dB SNR additive Gaussian noise. We developed a method to convert GEM-3 EM! measurements of unknown units into known quantities. This conversion factor was found through matching modeled responses of spheres to GEM-3 measurements of spheres. Recovery of a soil's susceptibility through GEM-3 measurements via our conversion factor provided validation of our findings. Our results will enables GEM-3 measurements of objects to be directly compared to modeled responses. Furthermore, the conversion factor will enable the magnetic properties of soil to be characterized through in situ measurement conditions.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 17, 2008
Accession Number
ADA500510

Entities

People

  • Bae-ian Wu
  • Beijia Zhang
  • Jin A. Kong
  • Kevin O’neill

Organizations

  • Massachusetts Institute of Technology

Tags

Communities of Interest

  • Advanced Electronics
  • Sensors

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Classification
  • Composite Materials
  • Coordinate Systems
  • Detectors
  • Gaussian Noise
  • Machine Learning
  • Magnetic Fields
  • Magnetic Properties
  • Materials
  • Measurement
  • Neural Networks
  • Physical Properties
  • Supervised Machine Learning
  • Training
  • Unexploded Ammunition

Readers

  • Image Processing and Computer Vision.
  • Military/Explosive Ordnance Disposal (EOD) Technology
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Neural Networks