A general framework for component substitution image fusion: An implementation using the fast image fusion method

  • Authors:

  • Wen Dou;Yunhao Chen;Xiaobing Li;Daniel Z. Sui

  • Affiliations:

  • Key Laboratory of Land Use, Ministry of Land and Resources, Beijing 100035, PR China and College of Resources Science and Technology, Beijing Normal University, Beijing 100875, PR China;College of Resources Science and Technology, Beijing Normal University, Beijing 100875, PR China;College of Resources Science and Technology, Beijing Normal University, Beijing 100875, PR China;Department of Geography, Texas A&M University, College Station, TX 77843-3147, USA

  • Venue:
  • Computers & Geosciences
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Many image fusion algorithms have been developed while more algorithms are being developed to improve the ability of preserving spectral information. Starting from the analysis of component substitution (COS) image fusion technique, a novel general component substitution (GCOS) image fusion framework is proposed, which could be used in three aspects: (1) comparative analysis to existing algorithms, (2) providing a fast technique for current COS fusion methods, and (3) guiding the development of the new COS algorithm. A demonstrative implementation of GCOS is provided, which can employ radiometric properties of sensors in the process of image fusion. An experiment based on degraded IKONOS images was carried out to demonstrate the effectiveness of the method, and the fusion image processed through the proposed method shows a higher correlation coefficient (CC) and a universal image quality index (UIQI), and a lower relative difference (RD) with the reference image in comparison to those yielded through Gram-Schmidt (GS) spectral sharpening, principal component analysis (PCA), smoothing filter-based intensity modulation (SFIM) and additive wavelet transform (AWT) methods, and provides more reasonable spatial details by visual validation. Validation of the proposed algorithm proved the ability of the proposed GCOS framework.