Advanced Radiometry for High Discrimination Explosive Fireball Discrimination

Abstract

The high explosive fireball phenomenological model for the mid wave infrared spectrum, developed by AFIT, performs classification from spectral signatures was modified to use radiometric intensities. Five bands were sequentially fit to derive the five physical fit parameters describing the fireball's temperature, size, soot absorption coefficient within 16% and emissions from the H2O and CO2 concentrations within 333% of the spectral model. This was improved by changing the model?s the band sizes, center, and fitting methods where all five fit parameters were matched to within 17% of spectral model. This demonstrated that a combination of radiometric intensities could be used in place of the spectral data. Interpreting the intensities into fit parameters provided and increased in classification capability with a Fisher Ratio (FR) =23 as opposed to a FR=4 when using the five raw intensities. A systematic search was performed to investigate classification potential using two, three and four radiometric bands combinations. The two-band search yielded a maximum FR of 6, a poor classification capability where the three and four-band search highlighted a highly confined spectral region centered at 5000cm-1 with a FR=41.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 01, 2009
Accession Number
ADA496108

Entities

People

  • Steven E. Slagle

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Counter IED
  • Sensors
  • Weapons Technologies

DTIC Thesaurus Topics

  • Absorption Coefficients
  • Air Force
  • Atmospheric Attenuation
  • Detection
  • Detectors
  • Explosive Devices
  • Explosives
  • High Explosives
  • Infrared Spectra
  • Intensity
  • Materials
  • Measurement
  • Pattern Recognition
  • Radiometry
  • Scattering
  • Spectra
  • Spectroscopy

Fields of Study

  • Physics

Readers

  • Combustion science or combustion engineering.
  • Image Processing and Computer Vision.
  • Regression Analysis.