Multi- and hyperspectral reconnaissance

Authors:

  • Thomas Svensson
  • Jörgen Ahlberg
  • Lars Allard
  • Svante Björklund
  • Leif Carlsson
  • Staffan Cronström
  • Jan Fagerström
  • Nils Karlsson
  • Roland Lindell
  • Ingmar Renhorn
  • Niclas Wadströmer

Publish date: 2009-01-08

Report number: FOI-R--2642--SE

Pages: 60

Written in: Swedish

Keywords:

  • Multispectral
  • hyperspectral
  • sensor simulation
  • sensor model
  • anomaly detection
  • information theory
  • interferometer
  • imaging spectrometry
  • target detection

Abstract

This report describes the project "Multi- and hyperspectral reconnaissance" that was performed during 2006-2008. The main question and concern in the project has been to investigate the ability of multi- and hyperspectral technology for detection and identification of difficult objects, e.g. camouflaged or partly hidden objects. The activities can be divided into four areas:  Sensor technology: Sensors och collecting of sensor data  Image analysis/target detection: Exploitation of sensor data and sensor information  Simulation of sensor- and atmosphere: Models for target detection  Sensor development: Future sensor systems The project has been part of the EDA project HIPOD. Key conclusions are that hyperspectral technology is a powerful technology for anomaly detection and target surveillance in daylight, which has been verified in the evaluation of mine data at FOI, and that the technology makes automation of the anomaly detection process possible. Extension into the thermal infrared would provide night capacity. Promising preliminary studies have been performed with a hyperspectral system based on interferometric technology, which has potentially good performance also in the dark.