Spectral Mixing of Camouflaged Targets.

Abstract

A fundamental problem in target detection is the separation of a target and its background, particularly when the target is camouflaged. It is possible to discern camouflaged objects in vegetative backgrounds using reflected light in the visible and infrared range. Reflectance data was taken of five camouflage nets draped over various vehicles with a predominately green background. The aim of this analysis was to reconstruct the spectrum of the observed scene using a linear combination of individual basis spectra called "pure" endmembers. Linear spectral mixing assumes that the observed spectral radiance may be modeled as a linear combination of members of a "pure" endmember spectral mixing library. The computer algorithm written for this analysis demonstrated the ability to use linear spectral mixing to reconstruct an observed spectrum. The analysis of the abundance mixtures showed that consistent exploitable patterns exist with this type of data. The task of reconstructing the observed spectra was performed with a crude, "non-pure" endmember library. Even greater success could be achieved with a more sophisticated and complete library. (MM)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1994
Accession Number
ADA293623

Entities

People

  • John W. Chandler
  • Suzanne E. Lyon

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Air Platforms
  • Biomedical
  • Energy and Power Technologies
  • Ground and Sea Platforms
  • Sensors

DTIC Thesaurus Topics

  • Atmospheric Attenuation
  • Cells
  • Cellular Structures
  • Computers
  • Detection
  • Detectors
  • Diffraction
  • Optical Properties
  • Optics
  • Reflectance
  • Refractive Index
  • Scattering
  • Second World War
  • Spectra
  • United States
  • United States Naval Academy
  • Visible Spectra

Readers

  • Atmospheric Science / Meteorology, specifically Wind Wave Turbulence.
  • Computer Vision.
  • Image Processing and Computer Vision.