Processing and fusion of thermal and video sequences for terrestrial long range observation systems

Leonid P. Yaroslavsky, Barak Fishbain, Alex Shteinman, Shai Gepshtein

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

A two stage fusion image processing system has been developed for visual and thermal range video streams. At the first stage, an intra-channel-inter- frame fusion is applied to each channel images independently, addressing each channel specific limitations. Noise removing processing by means of 3D (spatial-temporal) local DCT adaptive filtering is applied to the thermal channel video sequence. An algorithm that compensates image distortion due to the atmospheric turbulence is applied to the visual range channel video sequence. The second stage is interframe intra-channel fusion. At this stage, thermal and visual range channel image frames are fused frame-by-frame using a weighted average scheme with locally adapted weights. Experimental results obtained for both simulated and real life thermal and visual range image sequences are demonstrated.

Original languageEnglish
Title of host publicationProceedings of the Seventh International Conference on Information Fusion, FUSION 2004
EditorsP. Svensson, J. Schubert
Pages848-855
Number of pages8
StatePublished - 2004
EventProceedings of the Seventh International Conference on Information Fusion, FUSION 2004 - Stockholm, Sweden
Duration: 28 Jun 20041 Jul 2004

Publication series

NameProceedings of the Seventh International Conference on Information Fusion, FUSION 2004
Volume2

Conference

ConferenceProceedings of the Seventh International Conference on Information Fusion, FUSION 2004
Country/TerritorySweden
CityStockholm
Period28/06/041/07/04

Keywords

  • 3-D DCT
  • Elastic Registration Method
  • Fusion
  • Noise Estimation
  • Optical Flow
  • Sinc-interpolation
  • Weighted Average
  • Wiener Filtering

Fingerprint

Dive into the research topics of 'Processing and fusion of thermal and video sequences for terrestrial long range observation systems'. Together they form a unique fingerprint.

Cite this