Back to articles
Volume: 32 | Article ID: art00019
Color Restoration of Multispectral Images: Near-Infrared (NIR) filter-to-Color (RGB) image
  DOI :  10.2352/ISSN.2470-1173.2020.10.IPAS-180  Published OnlineJanuary 2020

Multispectral images captured by near-infrared (NIR) filtered devices have an attractive potential for numerous applications in computer vision, robot vision, and an artificial intelligence system. Although near-infrared (NIR) filtered devices display some unseen details, but their colors contain false colors affected main color components. Therefore, the color restoration of near-infrared (NIR) filtered images is a challenge for human perception. To overcome the challenge, the proposed method has four main steps: i) color balance with chromatic adaptation; ii) color component switching; iii) Hue-Saturation-Lightness adjustment; and iv) image fusion using a proportional-weighing metric. The proposed results achieve to: i) restore RGB color component, ii) remove the color-overlapped effect and generate structure details close to the ground truth images; and iii) preserve overall brightness.

Subject Areas :
Views 52
Downloads 14
 articleview.views 52
 articleview.downloads 14
  Cite this article 

THAWEESAK TRONGTIRAKUL, WERAPON CHIRACHARIT, SOS AGAIAN, "Color Restoration of Multispectral Images: Near-Infrared (NIR) filter-to-Color (RGB) imagein Proc. IS&T Int’l. Symp. on Electronic Imaging: Image Processing: Algorithms and Systems XVIII,  2020,  pp 180-1 - 180-6,

 Copy citation
  Copyright statement 
Copyright © Society for Imaging Science and Technology 2020
Electronic Imaging
Society for Imaging Science and Technology
7003 Kilworth Lane, Springfield, VA 22151 USA